1DBI::DBD(3) User Contributed Perl Documentation DBI::DBD(3)
2
6 DBI::DBD - Perl DBI Database Driver Writer's Guide
7
9 perldoc DBI::DBD
10 Version and volatility
12 This document is still a minimal draft which is in need of further
13 work.
14 The changes will occur both because the DBI specification is changing
16 and hence the requirements on DBD drivers change, and because feedback
17 from people reading this document will suggest improvements to it.
18 Please read the DBI documentation first and fully, including the DBI
20 FAQ. Then reread the DBI specification again as you're reading this.
21 It'll help.
22 This document is a patchwork of contributions from various authors.
24 More contributions (preferably as patches) are very welcome.
25
27 This document is primarily intended to help people writing new database
28 drivers for the Perl Database Interface (Perl DBI). It may also help
29 others interested in discovering why the internals of a DBD driver are
30 written the way they are.
31 This is a guide. Few (if any) of the statements in it are completely
33 authoritative under all possible circumstances. This means you will
34 need to use judgement in applying the guidelines in this document. If
35 in any doubt at all, please do contact the dbi-dev mailing list
36 (details given below) where Tim Bunce and other driver authors can
37 help.
38
40 The first rule for creating a new database driver for the Perl DBI is
41 very simple: DON'T!
42 There is usually a driver already available for the database you want
44 to use, almost regardless of which database you choose. Very often, the
45 database will provide an ODBC driver interface, so you can often use
46 DBD::ODBC to access the database. This is typically less convenient on
47 a Unix box than on a Microsoft Windows box, but there are numerous
48 options for ODBC driver managers on Unix too, and very often the ODBC
49 driver is provided by the database supplier.
50 Before deciding that you need to write a driver, do your homework to
52 ensure that you are not wasting your energies.
53 [As of December 2002, the consensus is that if you need an ODBC driver
55 manager on Unix, then the unixODBC driver (available from
56 <http://www.unixodbc.org/>) is the way to go.]
57 The second rule for creating a new database driver for the Perl DBI is
59 also very simple: Don't -- get someone else to do it for you!
60 Nevertheless, there are occasions when it is necessary to write a new
62 driver, often to use a proprietary language or API to access the
63 database more swiftly, or more comprehensively, than an ODBC driver
64 can. Then you should read this document very carefully, but with a
65 suitably sceptical eye.
66 If there is something in here that does not make any sense, question
68 it. You might be right that the information is bogus, but don't come
69 to that conclusion too quickly.
70 URLs and mailing lists
72 The primary web-site for locating DBI software and information is
73 http://dbi.perl.org/
75 There are two main and one auxiliary mailing lists for people working
77 with DBI. The primary lists are dbi-users@perl.org for general users
78 of DBI and DBD drivers, and dbi-dev@perl.org mainly for DBD driver
79 writers (don't join the dbi-dev list unless you have a good reason).
80 The auxiliary list is dbi-announce@perl.org for announcing new releases
81 of DBI or DBD drivers.
82 You can join these lists by accessing the web-site
84 <http://dbi.perl.org/>. The lists are closed so you cannot send email
85 to any of the lists unless you join the list first.
86 You should also consider monitoring the comp.lang.perl.* newsgroups,
88 especially comp.lang.perl.modules.
89 The Cheetah book
91 The definitive book on Perl DBI is the Cheetah book, so called because
92 of the picture on the cover. Its proper title is 'Programming the Perl
93 DBI: Database programming with Perl' by Alligator Descartes and Tim
94 Bunce, published by O'Reilly Associates, February 2000, ISBN
95 1-56592-699-4. Buy it now if you have not already done so, and read it.
96 Locating drivers
98 Before writing a new driver, it is in your interests to find out
99 whether there already is a driver for your database. If there is such
100 a driver, it would be much easier to make use of it than to write your
101 own!
102 The primary web-site for locating Perl software is
104 <http://search.cpan.org/>. You should look under the various modules
105 listings for the software you are after. For example:
106 http://search.cpan.org/modlist/Database_Interfaces
108 Follow the DBD:: and DBIx:: links at the top to see those subsets.
110 See the DBI docs for information on DBI web sites and mailing lists.
112 Registering a new driver
114 Before going through any official registration process, you will need
115 to establish that there is no driver already in the works. You'll do
116 that by asking the DBI mailing lists whether there is such a driver
117 available, or whether anybody is working on one.
118 When you get the go ahead, you will need to establish the name of the
120 driver and a prefix for the driver. Typically, the name is based on the
121 name of the database software it uses, and the prefix is a contraction
122 of that. Hence, DBD::Oracle has the name Oracle and the prefix 'ora_'.
123 The prefix must be lowercase and contain no underscores other than the
124 one at the end.
125 This information will be recorded in the DBI module. Apart from
127 documentation purposes, registration is a prerequisite for installing
128 private methods.
129 If you are writing a driver which will not be distributed on CPAN, then
131 you should choose a prefix beginning with 'x_', to avoid potential
132 prefix collisions with drivers registered in the future. Thus, if you
133 wrote a non-CPAN distributed driver called DBD::CustomDB, the prefix
134 might be 'x_cdb_'.
135 This document assumes you are writing a driver called DBD::Driver, and
137 that the prefix 'drv_' is assigned to the driver.
138 Two styles of database driver
140 There are two distinct styles of database driver that can be written to
141 work with the Perl DBI.
142 Your driver can be written in pure Perl, requiring no C compiler. When
144 feasible, this is the best solution, but most databases are not written
145 in such a way that this can be done. Some examples of pure Perl drivers
146 are DBD::File and DBD::CSV.
147 Alternatively, and most commonly, your driver will need to use some C
149 code to gain access to the database. This will be classified as a C/XS
150 driver.
151 What code will you write?
153 There are a number of files that need to be written for either a pure
154 Perl driver or a C/XS driver. There are no extra files needed only by a
155 pure Perl driver, but there are several extra files needed only by a
156 C/XS driver.
157 Files common to pure Perl and C/XS drivers
159 Assuming that your driver is called DBD::Driver, these files are:
161 · Makefile.PL
163 · META.yml
165 · README
167 · MANIFEST
169 · Driver.pm
171 · lib/Bundle/DBD/Driver.pm
173 · lib/DBD/Driver/Summary.pm
175 · t/*.t
177 The first four files are mandatory. Makefile.PL is used to control how
179 the driver is built and installed. The README file tells people who
180 download the file about how to build the module and any prerequisite
181 software that must be installed. The MANIFEST file is used by the
182 standard Perl module distribution mechanism. It lists all the source
183 files that need to be distributed with your module. Driver.pm is what
184 is loaded by the DBI code; it contains the methods peculiar to your
185 driver.
186 Although the META.yml file is not required you are advised to create
188 one. Of particular importance are the build_requires and
189 configure_requires attributes which newer CPAN modules understand. You
190 use these to tell the CPAN module (and CPANPLUS) that your build and
191 configure mechanisms require DBI. The best reference for META.yml (at
192 the time of writing) is
193 <http://module-build.sourceforge.net/META-spec-v1.4.html>. You can find
194 a reasonable example of a META.yml in DBD::ODBC.
195 The lib/Bundle/DBD/Driver.pm file allows you to specify other Perl
197 modules on which yours depends in a format that allows someone to type
198 a simple command and ensure that all the pre-requisites are in place as
199 well as building your driver.
200 The lib/DBD/Driver/Summary.pm file contains (an updated version of) the
202 information that was included - or that would have been included - in
203 the appendices of the Cheetah book as a summary of the abilities of
204 your driver and the associated database.
205 The files in the t subdirectory are unit tests for your driver. You
207 should write your tests as stringently as possible, while taking into
208 account the diversity of installations that you can encounter:
209 · Your tests should not casually modify operational databases.
211 · You should never damage existing tables in a database.
213 · You should code your tests to use a constrained name space within
215 the database. For example, the tables (and all other named objects)
216 that are created could all begin with 'dbd_drv_'.
217 · At the end of a test run, there should be no testing objects left
219 behind in the database.
220 · If you create any databases, you should remove them.
222 · If your database supports temporary tables that are automatically
224 removed at the end of a session, then exploit them as often as
225 possible.
226 · Try to make your tests independent of each other. If you have a
228 test t/t11dowhat.t that depends upon the successful running of
229 t/t10thingamy.t, people cannot run the single test case
230 t/t11dowhat.t. Further, running t/t11dowhat.t twice in a row is
231 likely to fail (at least, if t/t11dowhat.t modifies the database at
232 all) because the database at the start of the second run is not
233 what you saw at the start of the first run.
234 · Document in your README file what you do, and what privileges
236 people need to do it.
237 · You can, and probably should, sequence your tests by including a
239 test number before an abbreviated version of the test name; the
240 tests are run in the order in which the names are expanded by
241 shell-style globbing.
242 · It is in your interests to ensure that your tests work as widely as
244 possible.
245 Many drivers also install sub-modules DBD::Driver::SubModule for any of
247 a variety of different reasons, such as to support the metadata methods
248 (see the discussion of "METADATA METHODS" below). Such sub-modules are
249 conventionally stored in the directory lib/DBD/Driver. The module
250 itself would usually be in a file SubModule.pm. All such sub-modules
251 should themselves be version stamped (see the discussions far below).
252 Extra files needed by C/XS drivers
254 The software for a C/XS driver will typically contain at least four
256 extra files that are not relevant to a pure Perl driver.
257 · Driver.xs
259 · Driver.h
261 · dbdimp.h
263 · dbdimp.c
265 The Driver.xs file is used to generate C code that Perl can call to
267 gain access to the C functions you write that will, in turn, call down
268 onto your database software.
269 The Driver.h header is a stylized header that ensures you can access
271 the necessary Perl and DBI macros, types, and function declarations.
272 The dbdimp.h is used to specify which functions have been implemented
274 by your driver.
275 The dbdimp.c file is where you write the C code that does the real work
277 of translating between Perl-ish data types and what the database
278 expects to use and return.
279 There are some (mainly small, but very important) differences between
281 the contents of Makefile.PL and Driver.pm for pure Perl and C/XS
282 drivers, so those files are described both in the section on creating a
283 pure Perl driver and in the section on creating a C/XS driver.
284 Obviously, you can add extra source code files to the list.
286 Requirements on a driver and driver writer
288 To be remotely useful, your driver must be implemented in a format that
289 allows it to be distributed via CPAN, the Comprehensive Perl Archive
290 Network (<http://www.cpan.org/> and <http://search.cpan.org>). Of
291 course, it is easier if you do not have to meet this criterion, but you
292 will not be able to ask for much help if you do not do so, and no-one
293 is likely to want to install your module if they have to learn a new
294 installation mechanism.
295
297 Writing a pure Perl driver is surprisingly simple. However, there are
298 some problems you should be aware of. The best option is of course
299 picking up an existing driver and carefully modifying one method after
300 the other.
301 Also look carefully at DBD::AnyData and DBD::Template.
303 As an example we take a look at the DBD::File driver, a driver for
305 accessing plain files as tables, which is part of the DBD::CSV package.
306 The minimal set of files we have to implement are Makefile.PL, README,
308 MANIFEST and Driver.pm.
309 Pure Perl version of Makefile.PL
311 You typically start with writing Makefile.PL, a Makefile generator. The
312 contents of this file are described in detail in the
313 ExtUtils::MakeMaker man pages. It is definitely a good idea if you
314 start reading them. At least you should know about the variables
315 CONFIGURE, DEFINED, PM, DIR, EXE_FILES, INC, LIBS, LINKTYPE, NAME,
316 OPTIMIZE, PL_FILES, VERSION, VERSION_FROM, clean, depend, realclean
317 from the ExtUtils::MakeMaker man page: these are used in almost any
318 Makefile.PL.
319 Additionally read the section on Overriding MakeMaker Methods and the
321 descriptions of the distcheck, disttest and dist targets: They will
322 definitely be useful for you.
323 Of special importance for DBI drivers is the postamble method from the
325 ExtUtils::MM_Unix man page.
326 For Emacs users, I recommend the libscan method, which removes Emacs
328 backup files (file names which end with a tilde '~') from lists of
329 files.
330 Now an example, I use the word "Driver" wherever you should insert your
332 driver's name:
333 # -*- perl -*-
335 use ExtUtils::MakeMaker;
337 WriteMakefile(
339 dbd_edit_mm_attribs( {
340 'NAME' => 'DBD::Driver',
341 'VERSION_FROM' => 'Driver.pm',
342 'INC' => '',
343 'dist' => { 'SUFFIX' => '.gz',
344 'COMPRESS' => 'gzip -9f' },
345 'realclean' => { FILES => '*.xsi' },
346 'PREREQ_PM' => '1.03',
347 'CONFIGURE' => sub {
348 eval {require DBI::DBD;};
349 if ($@) {
350 warn $@;
351 exit 0;
352 }
353 my $dbi_arch_dir = dbd_dbi_arch_dir();
354 if (exists($opts{INC})) {
355 return {INC => "$opts{INC} -I$dbi_arch_dir"};
356 } else {
357 return {INC => "-I$dbi_arch_dir"};
358 }
359 }
360 },
361 { create_pp_tests => 1})
362 );
363 package MY;
365 sub postamble { return main::dbd_postamble(@_); }
366 sub libscan {
367 my ($self, $path) = @_;
368 ($path =~ m/\~$/) ? undef : $path;
369 }
370 Note the calls to "dbd_edit_mm_attribs()" and "dbd_postamble()".
372 The second hash reference in the call to "dbd_edit_mm_attribs()"
374 (containing "create_pp_tests()") is optional; you should not use it
375 unless your driver is a pure Perl driver (that is, it does not use C
376 and XS code). Therefore, the call to "dbd_edit_mm_attribs()" is not
377 relevant for C/XS drivers and may be omitted; simply use the (single)
378 hash reference containing NAME etc as the only argument to
379 "WriteMakefile()".
380 Note that the "dbd_edit_mm_attribs()" code will fail if you do not have
382 a t sub-directory containing at least one test case.
383 PREREQ_PM tells MakeMaker that DBI (version 1.03 in this case) is
385 required for this module. This will issue a warning that DBI 1.03 is
386 missing if someone attempts to install your DBD without DBI 1.03. See
387 CONFIGURE below for why this does not work reliably in stopping cpan
388 testers failing your module if DBI is not installed.
389 CONFIGURE is a subroutine called by MakeMaker during "WriteMakefile".
391 By putting the "require DBI::DBD" in this section we can attempt to
392 load DBI::DBD but if it is missing we exit with success. As we exit
393 successfully without creating a Makefile when DBI::DBD is missing cpan
394 testers will not report a failure. This may seem at odds with PREREQ_PM
395 but PREREQ_PM does not cause "WriteMakefile" to fail (unless you also
396 specify PREREQ_FATAL which is strongly discouraged by MakeMaker) so
397 "WriteMakefile" would continue to call "dbd_dbi_arch_dir" and fail.
398 All drivers must use "dbd_postamble()" or risk running into problems.
400 Note the specification of VERSION_FROM; the named file (Driver.pm) will
402 be scanned for the first line that looks like an assignment to
403 $VERSION, and the subsequent text will be used to determine the version
404 number. Note the commentary in ExtUtils::MakeMaker on the subject of
405 correctly formatted version numbers.
406 If your driver depends upon external software (it usually will), you
408 will need to add code to ensure that your environment is workable
409 before the call to "WriteMakefile()". If you need to check for the
410 existence of an external library and perhaps modify INC to include the
411 paths to where the external library header files are located and you
412 cannot find the library or header files make sure you output a message
413 saying they cannot be found but "exit 0" (success) before calling
414 "WriteMakefile" or CPAN testers will fail your module if the external
415 library is not found.
416 A full-fledged Makefile.PL can be quite large (for example, the files
418 for DBD::Oracle and DBD::Informix are both over 1000 lines long, and
419 the Informix one uses - and creates - auxiliary modules too).
420 See also ExtUtils::MakeMaker and ExtUtils::MM_Unix. Consider using
422 CPAN::MakeMaker in place of ExtUtils::MakeMaker.
423 README
425 The README file should describe what the driver is for, the pre-
426 requisites for the build process, the actual build process, how to
427 report errors, and who to report them to.
428 Users will find ways of breaking the driver build and test process
430 which you would never even have dreamed to be possible in your worst
431 nightmares. Therefore, you need to write this document defensively,
432 precisely and concisely.
433 As always, use the README from one of the established drivers as a
435 basis for your own; the version in DBD::Informix is worth a look as it
436 has been quite successful in heading off problems.
437 · Note that users will have versions of Perl and DBI that are both
439 older and newer than you expected, but this will seldom cause much
440 trouble. When it does, it will be because you are using features
441 of DBI that are not supported in the version they are using.
442 · Note that users will have versions of the database software that
444 are both older and newer than you expected. You will save yourself
445 time in the long run if you can identify the range of versions
446 which have been tested and warn about versions which are not known
447 to be OK.
448 · Note that many people trying to install your driver will not be
450 experts in the database software.
451 · Note that many people trying to install your driver will not be
453 experts in C or Perl.
454 MANIFEST
456 The MANIFEST will be used by the Makefile's dist target to build the
457 distribution tar file that is uploaded to CPAN. It should list every
458 file that you want to include in your distribution, one per line.
459 lib/Bundle/DBD/Driver.pm
461 The CPAN module provides an extremely powerful bundle mechanism that
462 allows you to specify pre-requisites for your driver.
463 The primary pre-requisite is Bundle::DBI; you may want or need to add
465 some more. With the bundle set up correctly, the user can type:
466 perl -MCPAN -e 'install Bundle::DBD::Driver'
468 and Perl will download, compile, test and install all the Perl modules
470 needed to build your driver.
471 The prerequisite modules are listed in the "CONTENTS" section, with the
473 official name of the module followed by a dash and an informal name or
474 description.
475 · Listing Bundle::DBI as the main pre-requisite simplifies life.
477 · Don't forget to list your driver.
479 · Note that unless the DBMS is itself a Perl module, you cannot list
481 it as a pre-requisite in this file.
482 · You should keep the version of the bundle the same as the version
484 of your driver.
485 · You should add configuration management, copyright, and licencing
487 information at the top.
488 A suitable skeleton for this file is shown below.
490 package Bundle::DBD::Driver;
492 $VERSION = '0.01';
494 1;
496 __END__
498 =head1 NAME
500 Bundle::DBD::Driver - A bundle to install all DBD::Driver related modules
502 =head1 SYNOPSIS
504 C<perl -MCPAN -e 'install Bundle::DBD::Driver'>
506 =head1 CONTENTS
508 Bundle::DBI - Bundle for DBI by TIMB (Tim Bunce)
510 DBD::Driver - DBD::Driver by YOU (Your Name)
512 =head1 DESCRIPTION
514 This bundle includes all the modules used by the Perl Database
516 Interface (DBI) driver for Driver (DBD::Driver), assuming the
517 use of DBI version 1.13 or later, created by Tim Bunce.
518 If you've not previously used the CPAN module to install any
520 bundles, you will be interrogated during its setup phase.
521 But when you've done it once, it remembers what you told it.
522 You could start by running:
523 C<perl -MCPAN -e 'install Bundle::CPAN'>
525 =head1 SEE ALSO
527 Bundle::DBI
529 =head1 AUTHOR
531 Your Name E<lt>F<you@yourdomain.com>E<gt>
533 =head1 THANKS
535 This bundle was created by ripping off Bundle::libnet created by
537 Graham Barr E<lt>F<gbarr@ti.com>E<gt>, and radically simplified
538 with some information from Jochen Wiedmann E<lt>F<joe@ispsoft.de>E<gt>.
539 The template was then included in the DBI::DBD documentation by
540 Jonathan Leffler E<lt>F<jleffler@informix.com>E<gt>.
541 =cut
543 lib/DBD/Driver/Summary.pm
545 There is no substitute for taking the summary file from a driver that
546 was documented in the Perl book (such as DBD::Oracle or DBD::Informix
547 or DBD::ODBC, to name but three), and adapting it to describe the
548 facilities available via DBD::Driver when accessing the Driver
549 database.
550 Pure Perl version of Driver.pm
552 The Driver.pm file defines the Perl module DBD::Driver for your driver.
553 It will define a package DBD::Driver along with some version
554 information, some variable definitions, and a function "driver()" which
555 will have a more or less standard structure.
556 It will also define three sub-packages of DBD::Driver:
558 DBD::Driver::dr
560 with methods "connect()", "data_sources()" and "disconnect_all()";
561 DBD::Driver::db
563 with methods such as "prepare()";
564 DBD::Driver::st
566 with methods such as "execute()" and "fetch()".
567 The Driver.pm file will also contain the documentation specific to
569 DBD::Driver in the format used by perldoc.
570 In a pure Perl driver, the Driver.pm file is the core of the
572 implementation. You will need to provide all the key methods needed by
573 DBI.
574 Now let's take a closer look at an excerpt of File.pm as an example.
576 We ignore things that are common to any module (even non-DBI modules)
577 or really specific to the DBD::File package.
578 The DBD::Driver package
580 The header
582 package DBD::File;
584 use strict;
586 use vars qw($VERSION $drh);
587 $VERSION = "1.23.00" # Version number of DBD::File
589 This is where the version number of your driver is specified, and is
591 where Makefile.PL looks for this information. Please ensure that any
592 other modules added with your driver are also version stamped so that
593 CPAN does not get confused.
594 It is recommended that you use a two-part (1.23) or three-part
596 (1.23.45) version number. Also consider the CPAN system, which gets
597 confused and considers version 1.10 to precede version 1.9, so that
598 using a raw CVS, RCS or SCCS version number is probably not appropriate
599 (despite being very common).
600 For Subversion you could use:
602 $VERSION = sprintf("12.%06d", q$Revision: 12345 $ =~ /(\d+)/o);
604 (use lots of leading zeros on the second portion so if you move the
606 code to a shared repository like svn.perl.org the much larger revision
607 numbers won't cause a problem, at least not for a few years). For RCS
608 or CVS you can use:
609 $VERSION = sprintf "%d.%02d", '$Revision: 11.21 $ ' =~ /(\d+)\.(\d+)/;
611 which pads out the fractional part with leading zeros so all is well
613 (so long as you don't go past x.99)
614 $drh = undef; # holds driver handle once initialized
616 This is where the driver handle will be stored, once created. Note
618 that you may assume there is only one handle for your driver.
619 The driver constructor
621 The "driver()" method is the driver handle constructor. Note that the
623 "driver()" method is in the DBD::Driver package, not in one of the sub-
624 packages DBD::Driver::dr, DBD::Driver::db, or DBD::Driver::db.
625 sub driver
627 {
628 return $drh if $drh; # already created - return same one
629 my ($class, $attr) = @_;
630 $class .= "::dr";
632 DBD::Driver::db->install_method('drv_example_dbh_method');
634 DBD::Driver::st->install_method('drv_example_sth_method');
635 # not a 'my' since we use it above to prevent multiple drivers
637 $drh = DBI::_new_drh($class, {
638 'Name' => 'File',
639 'Version' => $VERSION,
640 'Attribution' => 'DBD::File by Jochen Wiedmann',
641 })
642 or return undef;
643 return $drh;
645 }
646 This is a reasonable example of how DBI implements its handles. There
648 are three kinds: driver handles (typically stored in $drh; from now on
649 called drh or $drh), database handles (from now on called dbh or $dbh)
650 and statement handles (from now on called sth or $sth).
651 The prototype of "DBI::_new_drh()" is
653 $drh = DBI::_new_drh($class, $public_attrs, $private_attrs);
655 with the following arguments:
657 $class
659 is typically the class for your driver, (for example,
660 "DBD::File::dr"), passed as the first argument to the "driver()"
661 method.
662 $public_attrs
664 is a hash ref to attributes like Name, Version, and Attribution.
665 These are processed and used by DBI. You had better not make any
666 assumptions about them nor should you add private attributes here.
667 $private_attrs
669 This is another (optional) hash ref with your private attributes.
670 DBI will store them and otherwise leave them alone.
671 The "DBI::_new_drh()" method and the "driver()" method both return
673 "undef" for failure (in which case you must look at $DBI::err and
674 $DBI::errstr for the failure information, because you have no driver
675 handle to use).
676 Using install_method() to expose driver-private methods
678 DBD::Foo::db->install_method($method_name, \%attr);
680 Installs the driver-private method named by $method_name into the DBI
682 method dispatcher so it can be called directly, avoiding the need to
683 use the func() method.
684 It is called as a static method on the driver class to which the method
686 belongs. The method name must begin with the corresponding registered
687 driver-private prefix. For example, for DBD::Oracle $method_name must
688 being with '"ora_"', and for DBD::AnyData it must begin with '"ad_"'.
689 The "\%attr" attributes can be used to provide fine control over how
691 the DBI dispatcher handles the dispatching of the method. However it's
692 undocumented at the moment. See the IMA_* #define's in DBI.xs and the
693 O=>0x000x values in the initialization of %DBI::DBI_methods in DBI.pm.
694 (Volunteers to polish up and document the interface are very welcome to
695 get in touch via dbi-dev@perl.org).
696 Methods installed using install_method default to the standard error
698 handling behaviour for DBI methods: clearing err and errstr before
699 calling the method, and checking for errors to trigger RaiseError etc.
700 on return. This differs from the default behaviour of func().
701 Note for driver authors: The DBD::Foo::xx->install_method call won't
703 work until the class-hierarchy has been setup. Normally the DBI looks
704 after that just after the driver is loaded. This means install_method()
705 can't be called at the time the driver is loaded unless the class-
706 hierarchy is set up first. The way to do that is to call the
707 setup_driver() method:
708 DBI->setup_driver('DBD::Foo');
710 before using install_method().
712 The CLONE special subroutine
714 Also needed here, in the DBD::Driver package, is a "CLONE()" method
716 that will be called by perl when an interpreter is cloned. All your
717 "CLONE()" method needs to do, currently, is clear the cached $drh so
718 the new interpreter won't start using the cached $drh from the old
719 interpreter:
720 sub CLONE {
722 undef $drh;
723 }
724 See
726 <http://search.cpan.org/dist/perl/pod/perlmod.pod#Making_your_module_threadsafe>
727 for details.
728 The DBD::Driver::dr package
730 The next lines of code look as follows:
732 package DBD::Driver::dr; # ====== DRIVER ======
734 $DBD::Driver::dr::imp_data_size = 0;
736 Note that no @ISA is needed here, or for the other DBD::Driver::*
738 classes, because the DBI takes care of that for you when the driver is
739 loaded.
740 *FIX ME* Explain what the imp_data_size is, so that implementors aren't
742 practicing cargo-cult programming.
743 The database handle constructor
745 The database handle constructor is the driver's (hence the changed
747 namespace) "connect()" method:
748 sub connect
750 {
751 my ($drh, $dr_dsn, $user, $auth, $attr) = @_;
752 # Some database specific verifications, default settings
754 # and the like can go here. This should only include
755 # syntax checks or similar stuff where it's legal to
756 # 'die' in case of errors.
757 # For example, many database packages requires specific
758 # environment variables to be set; this could be where you
759 # validate that they are set, or default them if they are not set.
760 my $driver_prefix = "drv_"; # the assigned prefix for this driver
762 # Process attributes from the DSN; we assume ODBC syntax
764 # here, that is, the DSN looks like var1=val1;...;varN=valN
765 foreach my $var ( split /;/, $dr_dsn ) {
766 my ($attr_name, $attr_value) = split '=', $var, 2;
767 return $drh->set_err($DBI::stderr, "Can't parse DSN part '$var'")
768 unless defined $attr_value;
769 # add driver prefix to attribute name if it doesn't have it already
771 $attr_name = $driver_prefix.$attr_name
772 unless $attr_name =~ /^$driver_prefix/o;
773 # Store attribute into %$attr, replacing any existing value.
775 # The DBI will STORE() these into $dbh after we've connected
776 $attr->{$attr_name} = $attr_value;
777 }
778 # Get the attributes we'll use to connect.
780 # We use delete here because these no need to STORE them
781 my $db = delete $attr->{drv_database} || delete $attr->{drv_db}
782 or return $drh->set_err($DBI::stderr, "No database name given in DSN '$dr_dsn'");
783 my $host = delete $attr->{drv_host} || 'localhost';
784 my $port = delete $attr->{drv_port} || 123456;
785 # Assume you can attach to your database via drv_connect:
787 my $connection = drv_connect($db, $host, $port, $user, $auth)
788 or return $drh->set_err($DBI::stderr, "Can't connect to $dr_dsn: ...");
789 # create a 'blank' dbh (call superclass constructor)
791 my ($outer, $dbh) = DBI::_new_dbh($drh, { Name => $dr_dsn });
792 $dbh->STORE('Active', 1 );
794 $dbh->{drv_connection} = $connection;
795 return $outer;
797 }
798 This is mostly the same as in the driver handle constructor above. The
800 arguments are described in DBI.
801 The constructor "DBI::_new_dbh()" is called, returning a database
803 handle. The constructor's prototype is:
804 ($outer, $inner) = DBI::_new_dbh($drh, $public_attr, $private_attr);
806 with similar arguments to those in the driver handle constructor,
808 except that the $class is replaced by $drh. The Name attribute is a
809 standard DBI attribute (see "Database Handle Attributes" in DBI).
810 In scalar context, only the outer handle is returned.
812 Note the use of the "STORE()" method for setting the dbh attributes.
814 That's because within the driver code, the handle object you have is
815 the 'inner' handle of a tied hash, not the outer handle that the users
816 of your driver have.
817 Because you have the inner handle, tie magic doesn't get invoked when
819 you get or set values in the hash. This is often very handy for speed
820 when you want to get or set simple non-special driver-specific
821 attributes.
822 However, some attribute values, such as those handled by the DBI like
824 PrintError, don't actually exist in the hash and must be read via
825 "$h->FETCH($attrib)" and set via "$h->STORE($attrib, $value)". If in
826 any doubt, use these methods.
827 The data_sources() method
829 The "data_sources()" method must populate and return a list of valid
831 data sources, prefixed with the "dbi:Driver" incantation that allows
832 them to be used in the first argument of the "DBI->connect()" method.
833 An example of this might be scanning the $HOME/.odbcini file on Unix
834 for ODBC data sources (DSNs).
835 As a trivial example, consider a fixed list of data sources:
837 sub data_sources
839 {
840 my($drh, $attr) = @_;
841 my(@list) = ();
842 # You need more sophisticated code than this to set @list...
843 push @list, "dbi:Driver:abc";
844 push @list, "dbi:Driver:def";
845 push @list, "dbi:Driver:ghi";
846 # End of code to set @list
847 return @list;
848 }
849 The disconnect_all() method
851 If you need to release any resources when the driver is unloaded, you
853 can provide a disconnect_all method.
854 Other driver handle methods
856 If you need any other driver handle methods, they can follow here.
858 Error handling
860 It is quite likely that something fails in the connect method. With
862 DBD::File for example, you might catch an error when setting the
863 current directory to something not existent by using the (driver-
864 specific) f_dir attribute.
865 To report an error, you use the "set_err()" method:
867 $h->set_err($err, $errmsg, $state);
869 This will ensure that the error is recorded correctly and that
871 RaiseError and PrintError etc are handled correctly.
872 Typically you'll always use the method instance, aka your method's
874 first argument.
875 As "set_err()" always returns "undef" your error handling code can
877 usually be simplified to something like this:
878 return $h->set_err($err, $errmsg, $state) if ...;
880 The DBD::Driver::db package
882 package DBD::Driver::db; # ====== DATABASE ======
884 $DBD::Driver::db::imp_data_size = 0;
886 The statement handle constructor
888 There's nothing much new in the statement handle constructor, which is
890 the "prepare()" method:
891 sub prepare
893 {
894 my ($dbh, $statement, @attribs) = @_;
895 # create a 'blank' sth
897 my ($outer, $sth) = DBI::_new_sth($dbh, { Statement => $statement });
898 $sth->STORE('NUM_OF_PARAMS', ($statement =~ tr/?//));
900 $sth->{drv_params} = [];
902 return $outer;
904 }
905 This is still the same -- check the arguments and call the super class
907 constructor "DBI::_new_sth()". Again, in scalar context, only the outer
908 handle is returned. The Statement attribute should be cached as shown.
909 Note the prefix drv_ in the attribute names: it is required that all
911 your private attributes use a lowercase prefix unique to your driver.
912 As mentioned earlier in this document, the DBI contains a registry of
913 known driver prefixes and may one day warn about unknown attributes
914 that don't have a registered prefix.
915 Note that we parse the statement here in order to set the attribute
917 NUM_OF_PARAMS. The technique illustrated is not very reliable; it can
918 be confused by question marks appearing in quoted strings, delimited
919 identifiers or in SQL comments that are part of the SQL statement. We
920 could set NUM_OF_PARAMS in the "execute()" method instead because the
921 DBI specification explicitly allows a driver to defer this, but then
922 the user could not call "bind_param()".
923 Transaction handling
925 Pure Perl drivers will rarely support transactions. Thus your
927 "commit()" and "rollback()" methods will typically be quite simple:
928 sub commit
930 {
931 my ($dbh) = @_;
932 if ($dbh->FETCH('Warn')) {
933 warn("Commit ineffective while AutoCommit is on");
934 }
935 0;
936 }
937 sub rollback {
939 my ($dbh) = @_;
940 if ($dbh->FETCH('Warn')) {
941 warn("Rollback ineffective while AutoCommit is on");
942 }
943 0;
944 }
945 Or even simpler, just use the default methods provided by the DBI that
947 do nothing except return "undef".
948 The DBI's default "begin_work()" method can be used by inheritance.
950 The STORE() and FETCH() methods
952 These methods (that we have already used, see above) are called for
954 you, whenever the user does a:
955 $dbh->{$attr} = $val;
957 or, respectively,
959 $val = $dbh->{$attr};
961 See perltie for details on tied hash refs to understand why these
963 methods are required.
964 The DBI will handle most attributes for you, in particular attributes
966 like RaiseError or PrintError. All you have to do is handle your
967 driver's private attributes and any attributes, like AutoCommit and
968 ChopBlanks, that the DBI can't handle for you.
969 A good example might look like this:
971 sub STORE
973 {
974 my ($dbh, $attr, $val) = @_;
975 if ($attr eq 'AutoCommit') {
976 # AutoCommit is currently the only standard attribute we have
977 # to consider.
978 if (!$val) { die "Can't disable AutoCommit"; }
979 return 1;
980 }
981 if ($attr =~ m/^drv_/) {
982 # Handle only our private attributes here
983 # Note that we could trigger arbitrary actions.
984 # Ideally we should warn about unknown attributes.
985 $dbh->{$attr} = $val; # Yes, we are allowed to do this,
986 return 1; # but only for our private attributes
987 }
988 # Else pass up to DBI to handle for us
989 $dbh->SUPER::STORE($attr, $val);
990 }
991 sub FETCH
993 {
994 my ($dbh, $attr) = @_;
995 if ($attr eq 'AutoCommit') { return 1; }
996 if ($attr =~ m/^drv_/) {
997 # Handle only our private attributes here
998 # Note that we could trigger arbitrary actions.
999 return $dbh->{$attr}; # Yes, we are allowed to do this,
1000 # but only for our private attributes
1001 }
1002 # Else pass up to DBI to handle
1003 $dbh->SUPER::FETCH($attr);
1004 }
1005 The DBI will actually store and fetch driver-specific attributes (with
1007 all lowercase names) without warning or error, so there's actually no
1008 need to implement driver-specific any code in your "FETCH()" and
1009 "STORE()" methods unless you need extra logic/checks, beyond getting or
1010 setting the value.
1011 Unless your driver documentation indicates otherwise, the return value
1013 of the "STORE()" method is unspecified and the caller shouldn't use
1014 that value.
1015 Other database handle methods
1017 As with the driver package, other database handle methods may follow
1019 here. In particular you should consider a (possibly empty)
1020 "disconnect()" method and possibly a "quote()" method if DBI's default
1021 isn't correct for you. You may also need the "type_info_all()" and
1022 "get_info()" methods, as described elsewhere in this document.
1023 Where reasonable use "$h->SUPER::foo()" to call the DBI's method in
1025 some or all cases and just wrap your custom behavior around that.
1026 If you want to use private trace flags you'll probably want to be able
1028 to set them by name. To do that you'll need to define a
1029 "parse_trace_flag()" method (note that's "parse_trace_flag", singular,
1030 not "parse_trace_flags", plural).
1031 sub parse_trace_flag {
1033 my ($h, $name) = @_;
1034 return 0x01000000 if $name eq 'foo';
1035 return 0x02000000 if $name eq 'bar';
1036 return 0x04000000 if $name eq 'baz';
1037 return 0x08000000 if $name eq 'boo';
1038 return 0x10000000 if $name eq 'bop';
1039 return $h->SUPER::parse_trace_flag($name);
1040 }
1041 All private flag names must be lowercase, and all private flags must be
1043 in the top 8 of the 32 bits.
1044 The DBD::Driver::st package
1046 This package follows the same pattern the others do:
1048 package DBD::Driver::st;
1050 $DBD::Driver::st::imp_data_size = 0;
1052 The execute() and bind_param() methods
1054 This is perhaps the most difficult method because we have to consider
1056 parameter bindings here. In addition to that, there are a number of
1057 statement attributes which must be set for inherited DBI methods to
1058 function correctly (see "Statement attributes" below).
1059 We present a simplified implementation by using the drv_params
1061 attribute from above:
1062 sub bind_param
1064 {
1065 my ($sth, $pNum, $val, $attr) = @_;
1066 my $type = (ref $attr) ? $attr->{TYPE} : $attr;
1067 if ($type) {
1068 my $dbh = $sth->{Database};
1069 $val = $dbh->quote($sth, $type);
1070 }
1071 my $params = $sth->{drv_params};
1072 $params->[$pNum-1] = $val;
1073 1;
1074 }
1075 sub execute
1077 {
1078 my ($sth, @bind_values) = @_;
1079 # start of by finishing any previous execution if still active
1081 $sth->finish if $sth->FETCH('Active');
1082 my $params = (@bind_values) ?
1084 \@bind_values : $sth->{drv_params};
1085 my $numParam = $sth->FETCH('NUM_OF_PARAMS');
1086 return $sth->set_err($DBI::stderr, "Wrong number of parameters")
1087 if @$params != $numParam;
1088 my $statement = $sth->{'Statement'};
1089 for (my $i = 0; $i < $numParam; $i++) {
1090 $statement =~ s/?/$params->[$i]/; # XXX doesn't deal with quoting etc!
1091 }
1092 # Do anything ... we assume that an array ref of rows is
1093 # created and store it:
1094 $sth->{'drv_data'} = $data;
1095 $sth->{'drv_rows'} = @$data; # number of rows
1096 $sth->STORE('NUM_OF_FIELDS') = $numFields;
1097 $sth->{Active} = 1;
1098 @$data || '0E0';
1099 }
1100 There are a number of things you should note here.
1102 We initialize the NUM_OF_FIELDS and Active attributes here, because
1104 they are essential for "bind_columns()" to work.
1105 We use attribute "$sth->{Statement}" which we created within
1107 "prepare()". The attribute "$sth->{Database}", which is nothing else
1108 than the dbh, was automatically created by DBI.
1109 Finally, note that (as specified in the DBI specification) we return
1111 the string '0E0' instead of the number 0, so that the result tests true
1112 but equal to zero.
1113 $sth->execute() or die $sth->errstr;
1115 The execute_array(), execute_for_fetch() and bind_param_array() methods
1117 In general, DBD's only need to implement "execute_for_fetch()" and
1119 "bind_param_array". DBI's default "execute_array()" will invoke the
1120 DBD's "execute_for_fetch()" as needed.
1121 The following sequence describes the interaction between DBI
1123 "execute_array" and a DBD's "execute_for_fetch":
1124 1. App calls "$sth->execute_array(\%attrs, @array_of_arrays)"
1126 2. If @array_of_arrays was specified, DBI processes @array_of_arrays
1128 by calling DBD's "bind_param_array()". Alternately, App may have
1129 directly called "bind_param_array()"
1130 3. DBD validates and binds each array
1132 4. DBI retrieves the validated param arrays from DBD's ParamArray
1134 attribute
1135 5. DBI calls DBD's "execute_for_fetch($fetch_tuple_sub,
1137 \@tuple_status)", where &$fetch_tuple_sub is a closure to iterate
1138 over the returned ParamArray values, and "\@tuple_status" is an
1139 array to receive the disposition status of each tuple.
1140 6. DBD iteratively calls &$fetch_tuple_sub to retrieve parameter
1142 tuples to be added to its bulk database operation/request.
1143 7. when DBD reaches the limit of tuples it can handle in a single
1145 database operation/request, or the &$fetch_tuple_sub indicates no
1146 more tuples by returning undef, the DBD executes the bulk
1147 operation, and reports the disposition of each tuple in
1148 \@tuple_status.
1149 8. DBD repeats steps 6 and 7 until all tuples are processed.
1151 E.g., here's the essence of DBD::Oracle's execute_for_fetch:
1153 while (1) {
1155 my @tuple_batch;
1156 for (my $i = 0; $i < $batch_size; $i++) {
1157 push @tuple_batch, [ @{$fetch_tuple_sub->() || last} ];
1158 }
1159 last unless @tuple_batch;
1160 my $res = ora_execute_array($sth, \@tuple_batch,
1161 scalar(@tuple_batch), $tuple_batch_status);
1162 push @$tuple_status, @$tuple_batch_status;
1163 }
1164 Note that DBI's default execute_array()/execute_for_fetch()
1166 implementation requires the use of positional (i.e., '?') placeholders.
1167 Drivers which require named placeholders must either emulate positional
1168 placeholders (e.g., see DBD::Oracle), or must implement their own
1169 execute_array()/execute_for_fetch() methods to properly sequence bound
1170 parameter arrays.
1171 Fetching data
1173 Only one method needs to be written for fetching data,
1175 "fetchrow_arrayref()". The other methods, "fetchrow_array()",
1176 "fetchall_arrayref()", etc, as well as the database handle's "select*"
1177 methods are part of DBI, and call "fetchrow_arrayref()" as necessary.
1178 sub fetchrow_arrayref
1180 {
1181 my ($sth) = @_;
1182 my $data = $sth->{drv_data};
1183 my $row = shift @$data;
1184 if (!$row) {
1185 $sth->STORE(Active => 0); # mark as no longer active
1186 return undef;
1187 }
1188 if ($sth->FETCH('ChopBlanks')) {
1189 map { $_ =~ s/\s+$//; } @$row;
1190 }
1191 return $sth->_set_fbav($row);
1192 }
1193 *fetch = \&fetchrow_arrayref; # required alias for fetchrow_arrayref
1194 Note the use of the method "_set_fbav()" -- this is required so that
1196 "bind_col()" and "bind_columns()" work.
1197 If an error occurs which leaves the $sth in a state where remaining
1199 rows can't be fetched then Active should be turned off before the
1200 method returns.
1201 The "rows()" method for this driver can be implemented like this:
1203 sub rows { shift->{drv_rows} }
1205 because it knows in advance how many rows it has fetched.
1207 Alternatively you could delete that method and so fallback to the DBI's
1208 own method which does the right thing based on the number of calls to
1209 "_set_fbav()".
1210 The more_results method
1212 If your driver doesn't support multiple result sets, then don't even
1214 implement this method.
1215 Otherwise, this method needs to get the statement handle ready to fetch
1217 results from the next result set, if there is one. Typically you'd
1218 start with:
1219 $sth->finish;
1221 then you should delete all the attributes from the attribute cache that
1223 may no longer be relevant for the new result set:
1224 delete $sth->{$_}
1226 for qw(NAME TYPE PRECISION SCALE ...);
1227 for drivers written in C use:
1229 hv_delete((HV*)SvRV(sth), "NAME", 4, G_DISCARD);
1231 hv_delete((HV*)SvRV(sth), "NULLABLE", 8, G_DISCARD);
1232 hv_delete((HV*)SvRV(sth), "NUM_OF_FIELDS", 13, G_DISCARD);
1233 hv_delete((HV*)SvRV(sth), "PRECISION", 9, G_DISCARD);
1234 hv_delete((HV*)SvRV(sth), "SCALE", 5, G_DISCARD);
1235 hv_delete((HV*)SvRV(sth), "TYPE", 4, G_DISCARD);
1236 Don't forget to also delete, or update, any driver-private attributes
1238 that may not be correct for the next resultset.
1239 The NUM_OF_FIELDS attribute is a special case. It should be set using
1241 STORE:
1242 $sth->STORE(NUM_OF_FIELDS => 0); /* for DBI <= 1.53 */
1244 $sth->STORE(NUM_OF_FIELDS => $new_value);
1245 for drivers written in C use this incantation:
1247 /* Adjust NUM_OF_FIELDS - which also adjusts the row buffer size */
1249 DBIc_NUM_FIELDS(imp_sth) = 0; /* for DBI <= 1.53 */
1250 DBIc_STATE(imp_xxh)->set_attr_k(sth, sv_2mortal(newSVpvn("NUM_OF_FIELDS",13)), 0,
1251 sv_2mortal(newSViv(mysql_num_fields(imp_sth->result)))
1252 );
1253 For DBI versions prior to 1.54 you'll also need to explicitly adjust
1255 the number of elements in the row buffer array
1256 ("DBIc_FIELDS_AV(imp_sth)") to match the new result set. Fill any new
1257 values with newSV(0) not &sv_undef. Alternatively you could free
1258 DBIc_FIELDS_AV(imp_sth) and set it to null, but that would mean
1259 bind_columns() wouldn't work across result sets.
1260 Statement attributes
1262 The main difference between dbh and sth attributes is, that you should
1264 implement a lot of attributes here that are required by the DBI, such
1265 as NAME, NULLABLE, TYPE, etc. See "Statement Handle Attributes" in DBI
1266 for a complete list.
1267 Pay attention to attributes which are marked as read only, such as
1269 NUM_OF_PARAMS. These attributes can only be set the first time a
1270 statement is executed. If a statement is prepared, then executed
1271 multiple times, warnings may be generated.
1272 You can protect against these warnings, and prevent the recalculation
1274 of attributes which might be expensive to calculate (such as the NAME
1275 and NAME_* attributes):
1276 my $storedNumParams = $sth->FETCH('NUM_OF_PARAMS');
1278 if (!defined $storedNumParams or $storedNumFields < 0) {
1279 $sth->STORE('NUM_OF_PARAMS') = $numParams;
1280 # Set other useful attributes that only need to be set once
1282 # for a statement, like $sth->{NAME} and $sth->{TYPE}
1283 }
1284 One particularly important attribute to set correctly (mentioned in
1286 "ATTRIBUTES COMMON TO ALL HANDLES" in DBI is Active. Many DBI methods,
1287 including "bind_columns()", depend on this attribute.
1288 Besides that the "STORE()" and "FETCH()" methods are mainly the same as
1290 above for dbh's.
1291 Other statement methods
1293 A trivial "finish()" method to discard stored data, reset any
1295 attributes (such as Active) and do "$sth->SUPER::finish()".
1296 If you've defined a "parse_trace_flag()" method in ::db you'll also
1298 want it in ::st, so just alias it in:
1299 *parse_trace_flag = \&DBD::foo:db::parse_trace_flag;
1301 And perhaps some other methods that are not part of the DBI
1303 specification, in particular to make metadata available. Remember that
1304 they must have names that begin with your drivers registered prefix so
1305 they can be installed using "install_method()".
1306 If "DESTROY()" is called on a statement handle that's still active
1308 ("$sth->{Active}" is true) then it should effectively call "finish()".
1309 sub DESTROY {
1311 my $sth = shift;
1312 $sth->finish if $sth->FETCH('Active');
1313 }
1314 Tests
1316 The test process should conform as closely as possibly to the Perl
1317 standard test harness.
1318 In particular, most (all) of the tests should be run in the t sub-
1320 directory, and should simply produce an "ok" when run under "make
1321 test". For details on how this is done, see the Camel book and the
1322 section in Chapter 7, "The Standard Perl Library" on Test::Harness.
1323 The tests may need to adapt to the type of database which is being used
1325 for testing, and to the privileges of the user testing the driver. For
1326 example, the DBD::Informix test code has to adapt in a number of places
1327 to the type of database to which it is connected as different Informix
1328 databases have different capabilities: some of the tests are for
1329 databases without transaction logs; others are for databases with a
1330 transaction log; some versions of the server have support for blobs, or
1331 stored procedures, or user-defined data types, and others do not.
1332 When a complete file of tests must be skipped, you can provide a reason
1334 in a pseudo-comment:
1335 if ($no_transactions_available)
1337 {
1338 print "1..0 # Skip: No transactions available\n";
1339 exit 0;
1340 }
1341 Consider downloading the DBD::Informix code and look at the code in
1343 DBD/Informix/TestHarness.pm which is used throughout the DBD::Informix
1344 tests in the t sub-directory.
1345
1347 Please also see the section under "CREATING A PURE PERL DRIVER"
1348 regarding the creation of the Makefile.PL.
1349 Creating a new C/XS driver from scratch will always be a daunting task.
1351 You can and should greatly simplify your task by taking a good
1352 reference driver implementation and modifying that to match the
1353 database product for which you are writing a driver.
1354 The de facto reference driver has been the one for DBD::Oracle written
1356 by Tim Bunce, who is also the author of the DBI package. The
1357 DBD::Oracle module is a good example of a driver implemented around a
1358 C-level API.
1359 Nowadays it it seems better to base on DBD::ODBC, another driver
1361 maintained by Tim and Jeff Urlwin, because it offers a lot of metadata
1362 and seems to become the guideline for the future development. (Also as
1363 DBD::Oracle digs deeper into the Oracle 8 OCI interface it'll get even
1364 more hairy than it is now.)
1365 The DBD::Informix driver is one driver implemented using embedded SQL
1367 instead of a function-based API. DBD::Ingres may also be worth a look.
1368 C/XS version of Driver.pm
1370 A lot of the code in the Driver.pm file is very similar to the code for
1371 pure Perl modules - see above. However, there are also some subtle
1372 (and not so subtle) differences, including:
1373 · The variables $DBD::Driver::{dr|db|st}::imp_data_size are not
1375 defined here, but in the XS code, because they declare the size
1376 of certain C structures.
1377 · Some methods are typically moved to the XS code, in particular
1379 "prepare()", "execute()", "disconnect()", "disconnect_all()"
1380 and the "STORE()" and "FETCH()" methods.
1381 · Other methods are still part of Driver.pm, but have callbacks
1383 to the XS code.
1384 · If the driver-specific parts of the imp_drh_t structure need to
1386 be formally initialized (which does not seem to be a common
1387 requirement), then you need to add a call to an appropriate XS
1388 function in the driver method of "DBD::Driver::driver()", and
1389 you define the corresponding function in Driver.xs, and you
1390 define the C code in dbdimp.c and the prototype in dbdimp.h.
1391 For example, DBD::Informix has such a requirement, and adds the
1393 following call after the call to "_new_drh()" in Informix.pm:
1394 DBD::Informix::dr::driver_init($drh);
1396 and the following code in Informix.xs:
1398 # Initialize the DBD::Informix driver data structure
1400 void
1401 driver_init(drh)
1402 SV *drh
1403 CODE:
1404 ST(0) = dbd_ix_dr_driver_init(drh) ? &sv_yes : &sv_no;
1405 and the code in dbdimp.h declares:
1407 extern int dbd_ix_dr_driver_init(SV *drh);
1409 and the code in dbdimp.ec (equivalent to dbdimp.c) defines:
1411 /* Formally initialize the DBD::Informix driver structure */
1413 int
1414 dbd_ix_dr_driver(SV *drh)
1415 {
1416 D_imp_drh(drh);
1417 imp_drh->n_connections = 0; /* No active connections */
1418 imp_drh->current_connection = 0; /* No current connection */
1419 imp_drh->multipleconnections = (ESQLC_VERSION >= 600) ? True : False;
1420 dbd_ix_link_newhead(&imp_drh->head); /* Empty linked list of connections */
1421 return 1;
1422 }
1423 DBD::Oracle has a similar requirement but gets around it by
1425 checking whether the private data part of the driver handle is
1426 all zeroed out, rather than add extra functions.
1427 Now let's take a closer look at an excerpt from Oracle.pm (revised
1429 heavily to remove idiosyncrasies) as an example, ignoring things that
1430 were already discussed for pure Perl drivers.
1431 The connect method
1433 The connect method is the database handle constructor. You could write
1435 either of two versions of this method: either one which takes
1436 connection attributes (new code) and one which ignores them (old code
1437 only).
1438 If you ignore the connection attributes, then you omit all mention of
1440 the $auth variable (which is a reference to a hash of attributes), and
1441 the XS system manages the differences for you.
1442 sub connect
1444 {
1445 my ($drh, $dbname, $user, $auth, $attr) = @_;
1446 # Some database specific verifications, default settings
1448 # and the like following here. This should only include
1449 # syntax checks or similar stuff where it's legal to
1450 # 'die' in case of errors.
1451 my $dbh = DBI::_new_dbh($drh, {
1453 'Name' => $dbname,
1454 })
1455 or return undef;
1456 # Call the driver-specific function _login in Driver.xs file which
1458 # calls the DBMS-specific function(s) to connect to the database,
1459 # and populate internal handle data.
1460 DBD::Driver::db::_login($dbh, $dbname, $user, $auth, $attr)
1461 or return undef;
1462 $dbh;
1464 }
1465 This is mostly the same as in the pure Perl case, the exception being
1467 the use of the private "_login()" callback, which is the function that
1468 will really connect to the database. It is implemented in Driver.xst
1469 (you should not implement it) and calls "dbd_db_login6()" or
1470 "dbd_db_login6_sv" from dbdimp.c. See below for details.
1471 If your driver has driver-specific attributes which may be passed in
1473 the connect method and hence end up in $attr in "dbd_db_login6" then it
1474 is best to delete any you process so DBI does not send them again via
1475 STORE after connect. You can do this in C like this:
1476 DBD_ATTRIB_DELETE(attr, "my_attribute_name",
1478 strlen("my_attribute_name"));
1479 However, prior to DBI subversion version 11605 (and fixed post 1.607)
1481 DBD_ATTRIB_DELETE segfaulted so if you cannot guarantee the DBI version
1482 will be post 1.607 you need to use:
1483 hv_delete((HV*)SvRV(attr), "my_attribute_name",
1485 strlen("my_attribute_name"), G_DISCARD);
1486 *FIX ME* Discuss removing attributes in Perl code.
1488 The disconnect_all method
1490 *FIX ME* T.B.S
1492 The data_sources method
1494 If your "data_sources()" method can be implemented in pure Perl, then
1496 do so because it is easier than doing it in XS code (see the section
1497 above for pure Perl drivers).
1498 If your "data_sources()" method must call onto compiled functions, then
1500 you will need to define dbd_dr_data_sources in your dbdimp.h file,
1501 which will trigger Driver.xst (in DBI v1.33 or greater) to generate the
1502 XS code that calls your actual C function (see the discussion below for
1503 details) and you do not code anything in Driver.pm to handle it.
1504 The prepare method
1506 The prepare method is the statement handle constructor, and most of it
1508 is not new. Like the "connect()" method, it now has a C callback:
1509 package DBD::Driver::db; # ====== DATABASE ======
1511 use strict;
1512 sub prepare
1514 {
1515 my ($dbh, $statement, $attribs) = @_;
1516 # create a 'blank' sth
1518 my $sth = DBI::_new_sth($dbh, {
1519 'Statement' => $statement,
1520 })
1521 or return undef;
1522 # Call the driver-specific function _prepare in Driver.xs file
1524 # which calls the DBMS-specific function(s) to prepare a statement
1525 # and populate internal handle data.
1526 DBD::Driver::st::_prepare($sth, $statement, $attribs)
1527 or return undef;
1528 $sth;
1529 }
1530 The execute method
1532 *FIX ME* T.B.S
1534 The fetchrow_arrayref method
1536 *FIX ME* T.B.S
1538 Other methods?
1540 *FIX ME* T.B.S
1542 Driver.xs
1544 Driver.xs should look something like this:
1545 #include "Driver.h"
1547 DBISTATE_DECLARE;
1549 INCLUDE: Driver.xsi
1551 MODULE = DBD::Driver PACKAGE = DBD::Driver::dr
1553 /* Non-standard drh XS methods following here, if any. */
1555 /* If none (the usual case), omit the MODULE line above too. */
1556 MODULE = DBD::Driver PACKAGE = DBD::Driver::db
1558 /* Non-standard dbh XS methods following here, if any. */
1560 /* Currently this includes things like _list_tables from */
1561 /* DBD::mSQL and DBD::mysql. */
1562 MODULE = DBD::Driver PACKAGE = DBD::Driver::st
1564 /* Non-standard sth XS methods following here, if any. */
1566 /* In particular this includes things like _list_fields from */
1567 /* DBD::mSQL and DBD::mysql for accessing metadata. */
1568 Note especially the include of Driver.xsi here: DBI inserts stub
1570 functions for almost all private methods here which will typically do
1571 much work for you.
1572 Wherever you really have to implement something, it will call a private
1574 function in dbdimp.c, and this is what you have to implement.
1575 You need to set up an extra routine if your driver needs to export
1577 constants of its own, analogous to the SQL types available when you
1578 say:
1579 use DBI qw(:sql_types);
1581 *FIX ME* T.B.S
1583 Driver.h
1585 Driver.h is very simple and the operational contents should look like
1586 this:
1587 #ifndef DRIVER_H_INCLUDED
1589 #define DRIVER_H_INCLUDED
1590 #define NEED_DBIXS_VERSION 93 /* 93 for DBI versions 1.00 to 1.51+ */
1592 #define PERL_NO_GET_CONTEXT /* if used require DBI 1.51+ */
1593 #include <DBIXS.h> /* installed by the DBI module */
1595 #include "dbdimp.h"
1597 #include "dbivport.h" /* see below */
1599 #include <dbd_xsh.h> /* installed by the DBI module */
1601 #endif /* DRIVER_H_INCLUDED */
1603 The DBIXS.h header defines most of the interesting information that the
1605 writer of a driver needs.
1606 The file dbd_xsh.h header provides prototype declarations for the C
1608 functions that you might decide to implement. Note that you should
1609 normally only define one of "dbd_db_login()", "dbd_db_login6()" or
1610 "dbd_db_login6_sv" unless you are intent on supporting really old
1611 versions of DBI (prior to DBI 1.06) as well as modern versions. The
1612 only standard, DBI-mandated functions that you need write are those
1613 specified in the dbd_xsh.h header. You might also add extra driver-
1614 specific functions in Driver.xs.
1615 The dbivport.h file should be copied from the latest DBI release into
1617 your distribution each time you modify your driver. Its job is to allow
1618 you to enhance your code to work with the latest DBI API while still
1619 allowing your driver to be compiled and used with older versions of the
1620 DBI (for example, when the "DBIh_SET_ERR_CHAR()" macro was added to DBI
1621 1.41, an emulation of it was added to dbivport.h). This makes users
1622 happy and your life easier. Always read the notes in dbivport.h to
1623 check for any limitations in the emulation that you should be aware of.
1624 With DBI v1.51 or better I recommend that the driver defines
1626 PERL_NO_GET_CONTEXT before DBIXS.h is included. This can significantly
1627 improve efficiency when running under a thread enabled perl. (Remember
1628 that the standard perl in most Linux distributions is built with
1629 threads enabled. So is ActiveState perl for Windows, and perl built
1630 for Apache mod_perl2.) If you do this there are some things to keep in
1631 mind:
1632 · If PERL_NO_GET_CONTEXT is defined, then every function that calls
1634 the Perl API will need to start out with a "dTHX;" declaration.
1635 · You'll know which functions need this, because the C compiler will
1637 complain that the undeclared identifier "my_perl" is used if and
1638 only if the perl you are using to develop and test your driver has
1639 threads enabled.
1640 · If you don't remember to test with a thread-enabled perl before
1642 making a release it's likely that you'll get failure reports from
1643 users who are.
1644 · For driver private functions it is possible to gain even more
1646 efficiency by replacing "dTHX;" with "pTHX_" prepended to the
1647 parameter list and then "aTHX_" prepended to the argument list
1648 where the function is called.
1649 See "How multiple interpreters and concurrency are supported" in
1651 perlguts for additional information about PERL_NO_GET_CONTEXT.
1652 Implementation header dbdimp.h
1654 This header file has two jobs:
1655 First it defines data structures for your private part of the handles.
1657 Second it defines macros that rename the generic names like
1659 "dbd_db_login()" to database specific names like "ora_db_login()". This
1660 avoids name clashes and enables use of different drivers when you work
1661 with a statically linked perl.
1662 It also will have the important task of disabling XS methods that you
1664 don't want to implement.
1665 Finally, the macros will also be used to select alternate
1667 implementations of some functions. For example, the "dbd_db_login()"
1668 function is not passed the attribute hash.
1669 Since DBI v1.06, if a "dbd_db_login6()" macro is defined (for a
1671 function with 6 arguments), it will be used instead with the attribute
1672 hash passed as the sixth argument.
1673 Since DBI post v1.607, if a "dbd_db_login6_sv()" macro is defined (for
1675 a function like dbd_db_login6 but with scalar pointers for the dbname,
1676 username and password), it will be used instead. This will allow your
1677 login6 function to see if there are any Unicode characters in the
1678 dbname.
1679 People used to just pick Oracle's dbdimp.c and use the same names,
1681 structures and types. I strongly recommend against that. At first
1682 glance this saves time, but your implementation will be less readable.
1683 It was just hell when I had to separate DBI specific parts, Oracle
1684 specific parts, mSQL specific parts and mysql specific parts in
1685 DBD::mysql's dbdimp.h and dbdimp.c. (DBD::mysql was a port of DBD::mSQL
1686 which was based on DBD::Oracle.) [Seconded, based on the experience
1687 taking DBD::Informix apart, even though the version inherited in 1996
1688 was only based on DBD::Oracle.]
1689 This part of the driver is your exclusive part. Rewrite it from
1691 scratch, so it will be clean and short: in other words, a better piece
1692 of code. (Of course keep an eye on other people's work.)
1693 struct imp_drh_st {
1695 dbih_drc_t com; /* MUST be first element in structure */
1696 /* Insert your driver handle attributes here */
1697 };
1698 struct imp_dbh_st {
1700 dbih_dbc_t com; /* MUST be first element in structure */
1701 /* Insert your database handle attributes here */
1702 };
1703 struct imp_sth_st {
1705 dbih_stc_t com; /* MUST be first element in structure */
1706 /* Insert your statement handle attributes here */
1707 };
1708 /* Rename functions for avoiding name clashes; prototypes are */
1710 /* in dbd_xsh.h */
1711 #define dbd_init drv_dr_init
1712 #define dbd_db_login6_sv drv_db_login_sv
1713 #define dbd_db_do drv_db_do
1714 ... many more here ...
1715 These structures implement your private part of the handles.
1717 You have to use the name "imp_dbh_{dr|db|st}" and the first field must
1719 be of type dbih_drc_t|_dbc_t|_stc_t and must be called "com".
1720 You should never access these fields directly, except by using the
1722 DBIc_xxx() macros below.
1723 Implementation source dbdimp.c
1725 Conventionally, dbdimp.c is the main implementation file (but
1726 DBD::Informix calls the file dbdimp.ec). This section includes a short
1727 note on each function that is used in the Driver.xsi template and thus
1728 has to be implemented.
1729 Of course, you will probably also need to implement other support
1731 functions, which should usually be file static if they are placed in
1732 dbdimp.c. If they are placed in other files, you need to list those
1733 files in Makefile.PL (and MANIFEST) to handle them correctly.
1734 It is wise to adhere to a namespace convention for your functions to
1736 avoid conflicts. For example, for a driver with prefix drv_, you might
1737 call externally visible functions dbd_drv_xxxx. You should also avoid
1738 non-constant global variables as much as possible to improve the
1739 support for threading.
1740 Since Perl requires support for function prototypes (ANSI or ISO or
1742 Standard C), you should write your code using function prototypes too.
1743 It is possible to use either the unmapped names such as "dbd_init()" or
1745 the mapped names such as "dbd_ix_dr_init()" in the dbdimp.c file.
1746 DBD::Informix uses the mapped names which makes it easier to identify
1747 where to look for linkage problems at runtime (which will report errors
1748 using the mapped names).
1749 Most other drivers, and in particular DBD::Oracle, use the unmapped
1751 names in the source code which makes it a little easier to compare code
1752 between drivers and eases discussions on the dbi-dev mailing list. The
1753 majority of the code fragments here will use the unmapped names.
1754 Ultimately, you should provide implementations for most of the
1756 functions listed in the dbd_xsh.h header. The exceptions are optional
1757 functions (such as "dbd_st_rows()") and those functions with
1758 alternative signatures, such as "dbd_db_login6_sv", "dbd_db_login6()"
1759 and dbd_db_login(). Then you should only implement one of the
1760 alternatives, and generally the newer one of the alternatives.
1761 The dbd_init method
1763 #include "Driver.h"
1765 DBISTATE_DECLARE;
1767 void dbd_init(dbistate_t* dbistate)
1769 {
1770 DBISTATE_INIT; /* Initialize the DBI macros */
1771 }
1772 The "dbd_init()" function will be called when your driver is first
1774 loaded; the bootstrap command in "DBD::Driver::dr::driver()" triggers
1775 this, and the call is generated in the BOOT section of Driver.xst.
1776 These statements are needed to allow your driver to use the DBI macros.
1777 They will include your private header file dbdimp.h in turn. Note that
1778 DBISTATE_INIT requires the name of the argument to "dbd_init()" to be
1779 called "dbistate()".
1780 The dbd_drv_error method
1782 You need a function to record errors so DBI can access them properly.
1784 You can call it whatever you like, but we'll call it "dbd_drv_error()"
1785 here.
1786 The argument list depends on your database software; different systems
1788 provide different ways to get at error information.
1789 static void dbd_drv_error(SV *h, int rc, const char *what)
1791 {
1792 Note that h is a generic handle, may it be a driver handle, a database
1794 or a statement handle.
1795 D_imp_xxh(h);
1797 This macro will declare and initialize a variable imp_xxh with a
1799 pointer to your private handle pointer. You may cast this to to
1800 imp_drh_t, imp_dbh_t or imp_sth_t.
1801 To record the error correctly, equivalent to the "set_err()" method,
1803 use one of the "DBIh_SET_ERR_CHAR(...)" or "DBIh_SET_ERR_SV(...)"
1804 macros, which were added in DBI 1.41:
1805 DBIh_SET_ERR_SV(h, imp_xxh, err, errstr, state, method);
1807 DBIh_SET_ERR_CHAR(h, imp_xxh, err_c, err_i, errstr, state, method);
1808 For "DBIh_SET_ERR_SV" the err, errstr, state, and method parameters are
1810 "SV*" (use &sv_undef instead of NULL).
1811 For "DBIh_SET_ERR_CHAR" the err_c, errstr, state, method parameters are
1813 "char*".
1814 The err_i parameter is an "IV" that's used instead of err_c if err_c is
1816 "Null".
1817 The method parameter can be ignored.
1819 The "DBIh_SET_ERR_CHAR" macro is usually the simplest to use when you
1821 just have an integer error code and an error message string:
1822 DBIh_SET_ERR_CHAR(h, imp_xxh, Nullch, rc, what, Nullch, Nullch);
1824 As you can see, any parameters that aren't relevant to you can be
1826 "Null".
1827 To make drivers compatible with DBI < 1.41 you should be using
1829 dbivport.h as described in "Driver.h" above.
1830 The (obsolete) macros such as "DBIh_EVENT2" should be removed from
1832 drivers.
1833 The names "dbis" and "DBIS", which were used in previous versions of
1835 this document, should be replaced with the "DBIc_DBISTATE(imp_xxh)"
1836 macro.
1837 The name "DBILOGFP", which was also used in previous versions of this
1839 document, should be replaced by "DBIc_LOGPIO(imp_xxh)".
1840 Your code should not call the C "<stdio.h>" I/O functions; you should
1842 use "PerlIO_printf()" as shown:
1843 if (DBIc_TRACE_LEVEL(imp_xxh) >= 2)
1845 PerlIO_printf(DBIc_LOGPIO(imp_xxh), "foobar %s: %s\n",
1846 foo, neatsvpv(errstr,0));
1847 That's the first time we see how tracing works within a DBI driver.
1849 Make use of this as often as you can, but don't output anything at a
1850 trace level less than 3. Levels 1 and 2 are reserved for the DBI.
1851 You can define up to 8 private trace flags using the top 8 bits of
1853 "DBIc_TRACE_FLAGS(imp)", that is: 0xFF000000. See the
1854 "parse_trace_flag()" method elsewhere in this document.
1855 The dbd_dr_data_sources method
1857 This method is optional; the support for it was added in DBI v1.33.
1859 As noted in the discussion of Driver.pm, if the data sources can be
1861 determined by pure Perl code, do it that way. If, as in DBD::Informix,
1862 the information is obtained by a C function call, then you need to
1863 define a function that matches the prototype:
1864 extern AV *dbd_dr_data_sources(SV *drh, imp_drh_t *imp_drh, SV *attrs);
1866 An outline implementation for DBD::Informix follows, assuming that the
1868 "sqgetdbs()" function call shown will return up to 100 databases names,
1869 with the pointers to each name in the array dbsname and the name
1870 strings themselves being stores in dbsarea.
1871 AV *dbd_dr_data_sources(SV *drh, imp_drh_t *imp_drh, SV *attr)
1873 {
1874 int ndbs;
1875 int i;
1876 char *dbsname[100];
1877 char dbsarea[10000];
1878 AV *av = Nullav;
1879 if (sqgetdbs(&ndbs, dbsname, 100, dbsarea, sizeof(dbsarea)) == 0)
1881 {
1882 av = NewAV();
1883 av_extend(av, (I32)ndbs);
1884 sv_2mortal((SV *)av);
1885 for (i = 0; i < ndbs; i++)
1886 av_store(av, i, newSVpvf("dbi:Informix:%s", dbsname[i]));
1887 }
1888 return(av);
1889 }
1890 The actual DBD::Informix implementation has a number of extra lines of
1892 code, logs function entry and exit, reports the error from
1893 "sqgetdbs()", and uses "#define"'d constants for the array sizes.
1894 The dbd_db_login6 method
1896 int dbd_db_login6_sv(SV* dbh, imp_dbh_t* imp_dbh, SV* dbname,
1898 SV* user, SV* auth, SV *attr);
1899 or
1901 int dbd_db_login6(SV* dbh, imp_dbh_t* imp_dbh, char* dbname,
1903 char* user, char* auth, SV *attr);
1904 This function will really connect to the database. The argument dbh is
1906 the database handle. imp_dbh is the pointer to the handles private
1907 data, as is imp_xxx in "dbd_drv_error()" above. The arguments dbname,
1908 user, auth and attr correspond to the arguments of the driver handle's
1909 "connect()" method.
1910 You will quite often use database specific attributes here, that are
1912 specified in the DSN. I recommend you parse the DSN (using Perl) within
1913 the "connect()" method and pass the segments of the DSN via the
1914 attributes parameter through "_login()" to "dbd_db_login6()".
1915 Here's how you fetch them; as an example we use hostname attribute,
1917 which can be up to 12 characters long excluding null terminator:
1918 SV** svp;
1920 STRLEN len;
1921 char* hostname;
1922 if ( (svp = DBD_ATTRIB_GET_SVP(attr, "drv_hostname", 12)) && SvTRUE(*svp)) {
1924 hostname = SvPV(*svp, len);
1925 DBD_ATTRIB_DELETE(attr, "drv_hostname", 12); /* avoid later STORE */
1926 } else {
1927 hostname = "localhost";
1928 }
1929 If you handle any driver specific attributes in the dbd_db_login6
1931 method you probably want to delete them from "attr" (as above with
1932 DBD_ATTRIB_DELETE). If you don't delete your handled attributes DBI
1933 will call "STORE" for each attribute after the connect/login and this
1934 is at best redundant for attributes you have already processed.
1935 Note: Until revision 11605 (post DBI 1.607), there was a problem with
1937 DBD_ATTRIBUTE_DELETE so unless you require a DBI version after 1.607
1938 you need to replace each DBD_ATTRIBUTE_DELETE call with:
1939 hv_delete((HV*)SvRV(attr), key, key_len, G_DISCARD)
1941 Note that you can also obtain standard attributes such as AutoCommit
1943 and ChopBlanks from the attributes parameter, using "DBD_ATTRIB_GET_IV"
1944 for integer attributes.
1945 If, for example, your database does not support transactions but
1947 AutoCommit is set off (requesting transaction support), then you can
1948 emulate a 'failure to connect'.
1949 Now you should really connect to the database. In general, if the
1951 connection fails, it is best to ensure that all allocated resources are
1952 released so that the handle does not need to be destroyed separately.
1953 If you are successful (and possibly even if you fail but you have
1954 allocated some resources), you should use the following macros:
1955 DBIc_IMPSET_on(imp_dbh);
1957 This indicates that the driver (implementor) has allocated resources in
1959 the imp_dbh structure and that the implementors private
1960 "dbd_db_destroy()" function should be called when the handle is
1961 destroyed.
1962 DBIc_ACTIVE_on(imp_dbh);
1964 This indicates that the handle has an active connection to the server
1966 and that the "dbd_db_disconnect()" function should be called before the
1967 handle is destroyed.
1968 Note that if you do need to fail, you should report errors via the drh
1970 or imp_drh rather than via dbh or imp_dbh because imp_dbh will be
1971 destroyed by the failure, so errors recorded in that handle will not be
1972 visible to DBI, and hence not the user either.
1973 Note too, that the function is passed dbh and imp_dbh, and there is a
1975 macro "D_imp_drh_from_dbh" which can recover the imp_drh from the
1976 imp_dbh. However, there is no DBI macro to provide you with the drh
1977 given either the imp_dbh or the dbh or the imp_drh (and there's no way
1978 to recover the dbh given just the imp_dbh).
1979 This suggests that, despite the above notes about "dbd_drv_error()"
1981 taking an "SV *", it may be better to have two error routines, one
1982 taking imp_dbh and one taking imp_drh instead. With care, you can
1983 factor most of the formatting code out so that these are small routines
1984 calling a common error formatter. See the code in DBD::Informix 1.05.00
1985 for more information.
1986 The "dbd_db_login6()" function should return TRUE for success, FALSE
1988 otherwise.
1989 Drivers implemented long ago may define the five-argument function
1991 "dbd_db_login()" instead of "dbd_db_login6()". The missing argument is
1992 the attributes. There are ways to work around the missing attributes,
1993 but they are ungainly; it is much better to use the 6-argument form.
1994 Even later drivers will use "dbd_db_login6_sv()" which provides the
1995 dbname, username and password as SVs.
1996 The dbd_db_commit and dbd_db_rollback methods
1998 int dbd_db_commit(SV *dbh, imp_dbh_t *imp_dbh);
2000 int dbd_db_rollback(SV* dbh, imp_dbh_t* imp_dbh);
2001 These are used for commit and rollback. They should return TRUE for
2003 success, FALSE for error.
2004 The arguments dbh and imp_dbh are the same as for "dbd_db_login6()"
2006 above; I will omit describing them in what follows, as they appear
2007 always.
2008 These functions should return TRUE for success, FALSE otherwise.
2010 The dbd_db_disconnect method
2012 This is your private part of the "disconnect()" method. Any dbh with
2014 the ACTIVE flag on must be disconnected. (Note that you have to set it
2015 in "dbd_db_connect()" above.)
2016 int dbd_db_disconnect(SV* dbh, imp_dbh_t* imp_dbh);
2018 The database handle will return TRUE for success, FALSE otherwise. In
2020 any case it should do a:
2021 DBIc_ACTIVE_off(imp_dbh);
2023 before returning so DBI knows that "dbd_db_disconnect()" was executed.
2025 Note that there's nothing to stop a dbh being disconnected while it
2027 still have active children. If your database API reacts badly to trying
2028 to use an sth in this situation then you'll need to add code like this
2029 to all sth methods:
2030 if (!DBIc_ACTIVE(DBIc_PARENT_COM(imp_sth)))
2032 return 0;
2033 Alternatively, you can add code to your driver to keep explicit track
2035 of the statement handles that exist for each database handle and
2036 arrange to destroy those handles before disconnecting from the
2037 database. There is code to do this in DBD::Informix. Similar comments
2038 apply to the driver handle keeping track of all the database handles.
2039 Note that the code which destroys the subordinate handles should only
2041 release the associated database resources and mark the handles
2042 inactive; it does not attempt to free the actual handle structures.
2043 This function should return TRUE for success, FALSE otherwise, but it
2045 is not clear what anything can do about a failure.
2046 The dbd_db_discon_all method
2048 int dbd_discon_all (SV *drh, imp_drh_t *imp_drh);
2050 This function may be called at shutdown time. It should make best-
2052 efforts to disconnect all database handles - if possible. Some
2053 databases don't support that, in which case you can do nothing but
2054 return 'success'.
2055 This function should return TRUE for success, FALSE otherwise, but it
2057 is not clear what anything can do about a failure.
2058 The dbd_db_destroy method
2060 This is your private part of the database handle destructor. Any dbh
2062 with the IMPSET flag on must be destroyed, so that you can safely free
2063 resources. (Note that you have to set it in "dbd_db_connect()" above.)
2064 void dbd_db_destroy(SV* dbh, imp_dbh_t* imp_dbh)
2066 {
2067 DBIc_IMPSET_off(imp_dbh);
2068 }
2069 The DBI Driver.xst code will have called "dbd_db_disconnect()" for you,
2071 if the handle is still 'active', before calling "dbd_db_destroy()".
2072 Before returning the function must switch IMPSET to off, so DBI knows
2074 that the destructor was called.
2075 A DBI handle doesn't keep references to its children. But children do
2077 keep references to their parents. So a database handle won't be
2078 "DESTROY"'d until all its children have been "DESTROY"'d.
2079 The dbd_db_STORE_attrib method
2081 This function handles
2083 $dbh->{$key} = $value;
2085 Its prototype is:
2087 int dbd_db_STORE_attrib(SV* dbh, imp_dbh_t* imp_dbh, SV* keysv,
2089 SV* valuesv);
2090 You do not handle all attributes; on the contrary, you should not
2092 handle DBI attributes here: leave this to DBI. (There are two
2093 exceptions, AutoCommit and ChopBlanks, which you should care about.)
2094 The return value is TRUE if you have handled the attribute or FALSE
2096 otherwise. If you are handling an attribute and something fails, you
2097 should call "dbd_drv_error()", so DBI can raise exceptions, if desired.
2098 If "dbd_drv_error()" returns, however, you have a problem: the user
2099 will never know about the error, because he typically will not check
2100 "$dbh->errstr()".
2101 I cannot recommend a general way of going on, if "dbd_drv_error()"
2103 returns, but there are examples where even the DBI specification
2104 expects that you "croak()". (See the AutoCommit method in DBI.)
2105 If you have to store attributes, you should either use your private
2107 data structure imp_xxx, the handle hash (via "(HV*)SvRV(dbh)"), or use
2108 the private imp_data.
2109 The first is best for internal C values like integers or pointers and
2111 where speed is important within the driver. The handle hash is best for
2112 values the user may want to get/set via driver-specific attributes.
2113 The private imp_data is an additional "SV" attached to the handle. You
2114 could think of it as an unnamed handle attribute. It's not normally
2115 used.
2116 The dbd_db_FETCH_attrib method
2118 This is the counterpart of "dbd_db_STORE_attrib()", needed for:
2120 $value = $dbh->{$key};
2122 Its prototype is:
2124 SV* dbd_db_FETCH_attrib(SV* dbh, imp_dbh_t* imp_dbh, SV* keysv);
2126 Unlike all previous methods this returns an "SV" with the value. Note
2128 that you should normally execute "sv_2mortal()", if you return a
2129 nonconstant value. (Constant values are &sv_undef, &sv_no and &sv_yes.)
2130 Note, that DBI implements a caching algorithm for attribute values. If
2132 you think, that an attribute may be fetched, you store it in the dbh
2133 itself:
2134 if (cacheit) /* cache value for later DBI 'quick' fetch? */
2136 hv_store((HV*)SvRV(dbh), key, kl, cachesv, 0);
2137 The dbd_st_prepare method
2139 This is the private part of the "prepare()" method. Note that you must
2141 not really execute the statement here. You may, however, preparse and
2142 validate the statement, or do similar things.
2143 int dbd_st_prepare(SV* sth, imp_sth_t* imp_sth, char* statement,
2145 SV* attribs);
2146 A typical, simple, possibility is to do nothing and rely on the perl
2148 "prepare()" code that set the Statement attribute on the handle. This
2149 attribute can then be used by "dbd_st_execute()".
2150 If the driver supports placeholders then the NUM_OF_PARAMS attribute
2152 must be set correctly by "dbd_st_prepare()":
2153 DBIc_NUM_PARAMS(imp_sth) = ...
2155 If you can, you should also setup attributes like NUM_OF_FIELDS, NAME,
2157 etc. here, but DBI doesn't require that - they can be deferred until
2158 execute() is called. However, if you do, document it.
2159 In any case you should set the IMPSET flag, as you did in
2161 "dbd_db_connect()" above:
2162 DBIc_IMPSET_on(imp_sth);
2164 The dbd_st_execute method
2166 This is where a statement will really be executed.
2168 int dbd_st_execute(SV* sth, imp_sth_t* imp_sth);
2170 "dbd_st_execute" should return -2 for any error, -1 if the number of
2172 rows affected is unknown else it should be the number of affected
2173 (updated, inserted) rows.
2174 Note that you must be aware a statement may be executed repeatedly.
2176 Also, you should not expect that "finish()" will be called between two
2177 executions, so you might need code, like the following, near the start
2178 of the function:
2179 if (DBIc_ACTIVE(imp_sth))
2181 dbd_st_finish(h, imp_sth);
2182 If your driver supports the binding of parameters (it should!), but the
2184 database doesn't, you must do it here. This can be done as follows:
2185 SV *svp;
2187 char* statement = DBD_ATTRIB_GET_PV(h, "Statement", 9, svp, "");
2188 int numParam = DBIc_NUM_PARAMS(imp_sth);
2189 int i;
2190 for (i = 0; i < numParam; i++)
2192 {
2193 char* value = dbd_db_get_param(sth, imp_sth, i);
2194 /* It is your drivers task to implement dbd_db_get_param, */
2195 /* it must be setup as a counterpart of dbd_bind_ph. */
2196 /* Look for '?' and replace it with 'value'. Difficult */
2197 /* task, note that you may have question marks inside */
2198 /* quotes and comments the like ... :-( */
2199 /* See DBD::mysql for an example. (Don't look too deep into */
2200 /* the example, you will notice where I was lazy ...) */
2201 }
2202 The next thing is to really execute the statement.
2204 Note that you must set the attributes NUM_OF_FIELDS, NAME, etc when the
2206 statement is successfully executed if the driver has not already done
2207 so: they may be used even before a potential "fetchrow()". In
2208 particular you have to tell DBI the number of fields that the statement
2209 has, because it will be used by DBI internally. Thus the function will
2210 typically ends with:
2211 if (isSelectStatement) {
2213 DBIc_NUM_FIELDS(imp_sth) = numFields;
2214 DBIc_ACTIVE_on(imp_sth);
2215 }
2216 It is important that the ACTIVE flag only be set for "SELECT"
2218 statements (or any other statements that can return many values from
2219 the database using a cursor-like mechanism). See "dbd_db_connect()"
2220 above for more explanations.
2221 There plans for a preparse function to be provided by DBI, but this has
2223 not reached fruition yet. Meantime, if you want to know how ugly it
2224 can get, try looking at the "dbd_ix_preparse()" in DBD::Informix
2225 dbdimp.ec and the related functions in iustoken.c and sqltoken.c.
2226 The dbd_st_fetch method
2228 This function fetches a row of data. The row is stored in in an array,
2230 of "SV"'s that DBI prepares for you. This has two advantages: it is
2231 fast (you even reuse the "SV"'s, so they don't have to be created after
2232 the first "fetchrow()"), and it guarantees that DBI handles
2233 "bind_cols()" for you.
2234 What you do is the following:
2236 AV* av;
2238 int numFields = DBIc_NUM_FIELDS(imp_sth); /* Correct, if NUM_FIELDS
2239 is constant for this statement. There are drivers where this is
2240 not the case! */
2241 int chopBlanks = DBIc_is(imp_sth, DBIcf_ChopBlanks);
2242 int i;
2243 if (!fetch_new_row_of_data(...)) {
2245 ... /* check for error or end-of-data */
2246 DBIc_ACTIVE_off(imp_sth); /* turn off Active flag automatically */
2247 return Nullav;
2248 }
2249 /* get the fbav (field buffer array value) for this row */
2250 /* it is very important to only call this after you know */
2251 /* that you have a row of data to return. */
2252 av = DBIc_DBISTATE(imp_sth)->get_fbav(imp_sth);
2253 for (i = 0; i < numFields; i++) {
2254 SV* sv = fetch_a_field(..., i);
2255 if (chopBlanks && SvOK(sv) && type_is_blank_padded(field_type[i])) {
2256 /* Remove white space from end (only) of sv */
2257 }
2258 sv_setsv(AvARRAY(av)[i], sv); /* Note: (re)use! */
2259 }
2260 return av;
2261 There's no need to use a "fetch_a_field()" function returning an "SV*".
2263 It's more common to use your database API functions to fetch the data
2264 as character strings and use code like this:
2265 sv_setpvn(AvARRAY(av)[i], char_ptr, char_count);
2267 "NULL" values must be returned as "undef". You can use code like this:
2269 SvOK_off(AvARRAY(av)[i]);
2271 The function returns the "AV" prepared by DBI for success or "Nullav"
2273 otherwise.
2274 *FIX ME* Discuss what happens when there's no more data to fetch.
2276 Are errors permitted if another fetch occurs after the first fetch
2277 that reports no more data. (Permitted, not required.)
2278 If an error occurs which leaves the $sth in a state where remaining
2280 rows can't be fetched then Active should be turned off before the
2281 method returns.
2282 The dbd_st_finish3 method
2284 The "$sth->finish()" method can be called if the user wishes to
2286 indicate that no more rows will be fetched even if the database has
2287 more rows to offer, and the DBI code can call the function when handles
2288 are being destroyed. See the DBI specification for more background
2289 details.
2290 In both circumstances, the DBI code ends up calling the
2292 "dbd_st_finish3()" method (if you provide a mapping for
2293 "dbd_st_finish3()" in dbdimp.h), or "dbd_st_finish()" otherwise. The
2294 difference is that "dbd_st_finish3()" takes a third argument which is
2295 an "int" with the value 1 if it is being called from a "destroy()"
2296 method and 0 otherwise.
2297 Note that DBI v1.32 and earlier test on "dbd_db_finish3()" to call
2299 "dbd_st_finish3()"; if you provide "dbd_st_finish3()", either define
2300 "dbd_db_finish3()" too, or insist on DBI v1.33 or later.
2301 All it needs to do is turn off the Active flag for the sth. It will
2303 only be called by Driver.xst code, if the driver has set ACTIVE to on
2304 for the sth.
2305 Outline example:
2307 int dbd_st_finish3(SV* sth, imp_sth_t* imp_sth, int from_destroy) {
2309 if (DBIc_ACTIVE(imp_sth))
2310 {
2311 /* close cursor or equivalent action */
2312 DBIc_ACTIVE_off(imp_sth);
2313 }
2314 return 1;
2315 }
2316 The from_destroy parameter is true if "dbd_st_finish3()" is being
2318 called from "DESTROY()" - and so the statement is about to be
2319 destroyed. For many drivers there is no point in doing anything more
2320 than turning off the Active flag in this case.
2321 The function returns TRUE for success, FALSE otherwise, but there isn't
2323 a lot anyone can do to recover if there is an error.
2324 The dbd_st_destroy method
2326 This function is the private part of the statement handle destructor.
2328 void dbd_st_destroy(SV* sth, imp_sth_t* imp_sth) {
2330 ... /* any clean-up that's needed */
2331 DBIc_IMPSET_off(imp_sth); /* let DBI know we've done it */
2332 }
2333 The DBI Driver.xst code will call "dbd_st_finish()" for you, if the sth
2335 has the ACTIVE flag set, before calling "dbd_st_destroy()".
2336 The dbd_st_STORE_attrib and dbd_st_FETCH_attrib methods
2338 These functions correspond to "dbd_db_STORE()" and "dbd_db_FETCH()"
2340 attrib above, except that they are for statement handles. See above.
2341 int dbd_st_STORE_attrib(SV* sth, imp_sth_t* imp_sth, SV* keysv,
2343 SV* valuesv);
2344 SV* dbd_st_FETCH_attrib(SV* sth, imp_sth_t* imp_sth, SV* keysv);
2345 The dbd_bind_ph method
2347 This function is internally used by the "bind_param()" method, the
2349 "bind_param_inout()" method and by the DBI Driver.xst code if
2350 "execute()" is called with any bind parameters.
2351 int dbd_bind_ph (SV *sth, imp_sth_t *imp_sth, SV *param,
2353 SV *value, IV sql_type, SV *attribs,
2354 int is_inout, IV maxlen);
2355 The param argument holds an "IV" with the parameter number (1, 2, ...).
2357 The value argument is the parameter value and sql_type is its type.
2358 If your driver does not support "bind_param_inout()" then you should
2360 ignore maxlen and croak if is_inout is TRUE.
2361 If your driver does support "bind_param_inout()" then you should note
2363 that value is the "SV" after dereferencing the reference passed to
2364 "bind_param_inout()".
2365 In drivers of simple databases the function will, for example, store
2367 the value in a parameter array and use it later in "dbd_st_execute()".
2368 See the DBD::mysql driver for an example.
2369 Implementing bind_param_inout support
2371 To provide support for parameters bound by reference rather than by
2373 value, the driver must do a number of things. First, and most
2374 importantly, it must note the references and stash them in its own
2375 driver structure. Secondly, when a value is bound to a column, the
2376 driver must discard any previous reference bound to the column. On
2377 each execute, the driver must evaluate the references and internally
2378 bind the values resulting from the references. This is only applicable
2379 if the user writes:
2380 $sth->execute;
2382 If the user writes:
2384 $sth->execute(@values);
2386 then DBI automatically calls the binding code for each element of
2388 @values. These calls are indistinguishable from explicit user calls to
2389 "bind_param()".
2390 C/XS version of Makefile.PL
2392 The Makefile.PL file for a C/XS driver is similar to the code needed
2393 for a pure Perl driver, but there are a number of extra bits of
2394 information needed by the build system.
2395 For example, the attributes list passed to "WriteMakefile()" needs to
2397 specify the object files that need to be compiled and built into the
2398 shared object (DLL). This is often, but not necessarily, just dbdimp.o
2399 (unless that should be dbdimp.obj because you're building on MS
2400 Windows).
2401 Note that you can reliably determine the extension of the object files
2403 from the $Config{obj_ext} values, and there are many other useful
2404 pieces of configuration information lurking in that hash. You get
2405 access to it with:
2406 use Config;
2408 Methods which do not need to be written
2410 The DBI code implements the majority of the methods which are accessed
2411 using the notation "DBI->function()", the only exceptions being
2412 "DBI->connect()" and "DBI->data_sources()" which require support from
2413 the driver.
2414 The DBI code implements the following documented driver, database and
2416 statement functions which do not need to be written by the DBD driver
2417 writer.
2418 $dbh->do()
2420 The default implementation of this function prepares, executes and
2421 destroys the statement. This can be replaced if there is a better
2422 way to implement this, such as "EXECUTE IMMEDIATE" which can
2423 sometimes be used if there are no parameters.
2424 $h->errstr()
2426 $h->err()
2427 $h->state()
2428 $h->trace()
2429 The DBD driver does not need to worry about these routines at all.
2430 $h->{ChopBlanks}
2432 This attribute needs to be honored during "fetch()" operations, but
2433 does not need to be handled by the attribute handling code.
2434 $h->{RaiseError}
2436 The DBD driver does not need to worry about this attribute at all.
2437 $h->{PrintError}
2439 The DBD driver does not need to worry about this attribute at all.
2440 $sth->bind_col()
2442 Assuming the driver uses the "DBIc_DBISTATE(imp_xxh)->get_fbav()"
2443 function (C drivers, see below), or the "$sth->_set_fbav($data)"
2444 method (Perl drivers) the driver does not need to do anything about
2445 this routine.
2446 $sth->bind_columns()
2448 Regardless of whether the driver uses
2449 "DBIc_DBISTATE(imp_xxh)->get_fbav()", the driver does not need to
2450 do anything about this routine as it simply iteratively calls
2451 "$sth->bind_col()".
2452 The DBI code implements a default implementation of the following
2454 functions which do not need to be written by the DBD driver writer
2455 unless the default implementation is incorrect for the Driver.
2456 $dbh->quote()
2458 This should only be written if the database does not accept the
2459 ANSI SQL standard for quoting strings, with the string enclosed in
2460 single quotes and any embedded single quotes replaced by two
2461 consecutive single quotes.
2462 For the two argument form of quote, you need to implement the
2464 "type_info()" method to provide the information that quote needs.
2465 $dbh->ping()
2467 This should be implemented as a simple efficient way to determine
2468 whether the connection to the database is still alive. Typically
2469 code like this:
2470 sub ping {
2472 my $dbh = shift;
2473 $sth = $dbh->prepare_cached(q{
2474 select * from A_TABLE_NAME where 1=0
2475 }) or return 0;
2476 $sth->execute or return 0;
2477 $sth->finish;
2478 return 1;
2479 }
2480 where A_TABLE_NAME is the name of a table that always exists (such
2482 as a database system catalogue).
2483 $drh->default_user
2485 The default implementation of default_user will get the database
2486 username and password fields from $ENV{DBI_USER} and
2487 $ENV{DBI_PASS}. You can override this method. It is called as
2488 follows:
2489 ($user, $pass) = $drh->default_user($user, $pass, $attr)
2491
2493 The exposition above ignores the DBI MetaData methods. The metadata
2494 methods are all associated with a database handle.
2495 Using DBI::DBD::Metadata
2497 The DBI::DBD::Metadata module is a good semi-automatic way for the
2498 developer of a DBD module to write the "get_info()" and "type_info()"
2499 functions quickly and accurately.
2500 Generating the get_info method
2502 Prior to DBI v1.33, this existed as the method "write_getinfo_pm()" in
2504 the DBI::DBD module. From DBI v1.33, it exists as the method
2505 "write_getinfo_pm()" in the DBI::DBD::Metadata module. This discussion
2506 assumes you have DBI v1.33 or later.
2507 You examine the documentation for "write_getinfo_pm()" using:
2509 perldoc DBI::DBD::Metadata
2511 To use it, you need a Perl DBI driver for your database which
2513 implements the "get_info()" method. In practice, this means you need to
2514 install DBD::ODBC, an ODBC driver manager, and an ODBC driver for your
2515 database.
2516 With the pre-requisites in place, you might type:
2518 perl -MDBI::DBD::Metadata -we \
2520 "write_getinfo_pm (qw{ dbi:ODBC:foo_db username password Driver })"
2521 The procedure writes to standard output the code that should be added
2523 to your Driver.pm file and the code that should be written to
2524 lib/DBD/Driver/GetInfo.pm.
2525 You should review the output to ensure that it is sensible.
2527 Generating the type_info method
2529 Given the idea of the "write_getinfo_pm()" method, it was not hard to
2531 devise a parallel method, "write_typeinfo_pm()", which does the
2532 analogous job for the DBI "type_info_all()" metadata method. The
2533 "write_typeinfo_pm()" method was added to DBI v1.33.
2534 You examine the documentation for "write_typeinfo_pm()" using:
2536 perldoc DBI::DBD::Metadata
2538 The setup is exactly analogous to the mechanism described in
2540 "Generating the get_info method".
2541 With the pre-requisites in place, you might type:
2543 perl -MDBI::DBD::Metadata -we \
2545 "write_typeinfo (qw{ dbi:ODBC:foo_db username password Driver })"
2546 The procedure writes to standard output the code that should be added
2548 to your Driver.pm file and the code that should be written to
2549 lib/DBD/Driver/TypeInfo.pm.
2550 You should review the output to ensure that it is sensible.
2552 Writing DBD::Driver::db::get_info
2554 If you use the DBI::DBD::Metadata module, then the code you need is
2555 generated for you.
2556 If you decide not to use the DBI::DBD::Metadata module, you should
2558 probably borrow the code from a driver that has done so (eg
2559 DBD::Informix from version 1.05 onwards) and crib the code from there,
2560 or look at the code that generates that module and follow that. The
2561 method in Driver.pm will be very simple; the method in
2562 lib/DBD/Driver/GetInfo.pm is not very much more complex unless your
2563 DBMS itself is much more complex.
2564 Note that some of the DBI utility methods rely on information from the
2566 "get_info()" method to perform their operations correctly. See, for
2567 example, the "quote_identifier()" and quote methods, discussed below.
2568 Writing DBD::Driver::db::type_info_all
2570 If you use the "DBI::DBD::Metadata" module, then the code you need is
2571 generated for you.
2572 If you decide not to use the "DBI::DBD::Metadata" module, you should
2574 probably borrow the code from a driver that has done so (eg
2575 "DBD::Informix" from version 1.05 onwards) and crib the code from
2576 there, or look at the code that generates that module and follow that.
2577 The method in Driver.pm will be very simple; the method in
2578 lib/DBD/Driver/TypeInfo.pm is not very much more complex unless your
2579 DBMS itself is much more complex.
2580 Writing DBD::Driver::db::type_info
2582 The guidelines on writing this method are still not really clear. No
2583 sample implementation is available.
2584 Writing DBD::Driver::db::table_info
2586 *FIX ME* The guidelines on writing this method have not been written yet.
2587 No sample implementation is available.
2588 Writing DBD::Driver::db::column_info
2590 *FIX ME* The guidelines on writing this method have not been written yet.
2591 No sample implementation is available.
2592 Writing DBD::Driver::db::primary_key_info
2594 *FIX ME* The guidelines on writing this method have not been written yet.
2595 No sample implementation is available.
2596 Writing DBD::Driver::db::primary_key
2598 *FIX ME* The guidelines on writing this method have not been written yet.
2599 No sample implementation is available.
2600 Writing DBD::Driver::db::foreign_key_info
2602 *FIX ME* The guidelines on writing this method have not been written yet.
2603 No sample implementation is available.
2604 Writing DBD::Driver::db::tables
2606 This method generates an array of names in a format suitable for being
2607 embedded in SQL statements in places where a table name is expected.
2608 If your database hews close enough to the SQL standard or if you have
2610 implemented an appropriate "table_info()" function and and the
2611 appropriate "quote_identifier()" function, then the DBI default version
2612 of this method will work for your driver too.
2613 Otherwise, you have to write a function yourself, such as:
2615 sub tables
2617 {
2618 my($dbh, $cat, $sch, $tab, $typ) = @_;
2619 my(@res);
2620 my($sth) = $dbh->table_info($cat, $sch, $tab, $typ);
2621 my(@arr);
2622 while (@arr = $sth->fetchrow_array)
2623 {
2624 push @res, $dbh->quote_identifier($arr[0], $arr[1], $arr[2]);
2625 }
2626 return @res;
2627 }
2628 See also the default implementation in DBI.pm.
2630 Writing DBD::Driver::db::quote
2632 This method takes a value and converts it into a string suitable for
2633 embedding in an SQL statement as a string literal.
2634 If your DBMS accepts the SQL standard notation for strings (single
2636 quotes around the string as a whole with any embedded single quotes
2637 doubled up), then you do not need to write this method as DBI provides
2638 a default method that does it for you.
2639 If your DBMS uses an alternative notation or escape mechanism, then you
2641 need to provide an equivalent function. For example, suppose your DBMS
2642 used C notation with double quotes around the string and backslashes
2643 escaping both double quotes and backslashes themselves. Then you might
2644 write the function as:
2645 sub quote
2647 {
2648 my($dbh, $str) = @_;
2649 $str =~ s/["\\]/\\$&/gmo;
2650 return qq{"$str"};
2651 }
2652 Handling newlines and other control characters is left as an exercise
2654 for the reader.
2655 This sample method ignores the $data_type indicator which is the
2657 optional second argument to the method.
2658 Writing DBD::Driver::db::quote_identifier
2660 This method is called to ensure that the name of the given table (or
2661 other database object) can be embedded into an SQL statement without
2662 danger of misinterpretation. The result string should be usable in the
2663 text of an SQL statement as the identifier for a table.
2664 If your DBMS accepts the SQL standard notation for quoted identifiers
2666 (which uses double quotes around the identifier as a whole, with any
2667 embedded double quotes doubled up) and accepts "schema"."identifier"
2668 (and "catalog"."schema"."identifier" when a catalog is specified), then
2669 you do not need to write this method as DBI provides a default method
2670 that does it for you.
2671 In fact, even if your DBMS does not handle exactly that notation but
2673 you have implemented the "get_info()" method and it gives the correct
2674 responses, then it will work for you. If your database is fussier, then
2675 you need to implement your own version of the function.
2676 For example, DBD::Informix has to deal with an environment variable
2678 DELIMIDENT. If it is not set, then the DBMS treats names enclosed in
2679 double quotes as strings rather than names, which is usually a syntax
2680 error. Additionally, the catalog portion of the name is separated from
2681 the schema and table by a different delimiter (colon instead of dot),
2682 and the catalog portion is never enclosed in quotes. (Fortunately,
2683 valid strings for the catalog will never contain weird characters that
2684 might need to be escaped, unless you count dots, dashes, slashes and
2685 at-signs as weird.) Finally, an Informix database can contain objects
2686 that cannot be accessed because they were created by a user with the
2687 DELIMIDENT environment variable set, but the current user does not have
2688 it set. By design choice, the "quote_identifier()" method encloses
2689 those identifiers in double quotes anyway, which generally triggers a
2690 syntax error, and the metadata methods which generate lists of tables
2691 etc omit those identifiers from the result sets.
2692 sub quote_identifier
2694 {
2695 my($dbh, $cat, $sch, $obj) = @_;
2696 my($rv) = "";
2697 my($qq) = (defined $ENV{DELIMIDENT}) ? '"' : '';
2698 $rv .= qq{$cat:} if (defined $cat);
2699 if (defined $sch)
2700 {
2701 if ($sch !~ m/^\w+$/o)
2702 {
2703 $qq = '"';
2704 $sch =~ s/$qq/$qq$qq/gm;
2705 }
2706 $rv .= qq{$qq$sch$qq.};
2707 }
2708 if (defined $obj)
2709 {
2710 if ($obj !~ m/^\w+$/o)
2711 {
2712 $qq = '"';
2713 $obj =~ s/$qq/$qq$qq/gm;
2714 }
2715 $rv .= qq{$qq$obj$qq};
2716 }
2717 return $rv;
2718 }
2719 Handling newlines and other control characters is left as an exercise
2721 for the reader.
2722 Note that there is an optional fourth parameter to this function which
2724 is a reference to a hash of attributes; this sample implementation
2725 ignores that.
2726 This sample implementation also ignores the single-argument variant of
2728 the method.
2729
2731 Tracing in DBI is controlled with a combination of a trace level and a
2732 set of flags which together are known as the trace settings. The trace
2733 settings are stored in a single integer and divided into levels and
2734 flags by a set of masks ("DBIc_TRACE_LEVEL_MASK" and
2735 "DBIc_TRACE_FLAGS_MASK").
2736 Each handle has it's own trace settings and so does the DBI. When you
2738 call a method the DBI merges the handles settings into its own for the
2739 duration of the call: the trace flags of the handle are OR'd into the
2740 trace flags of the DBI, and if the handle has a higher trace level then
2741 the DBI trace level is raised to match it. The previous DBI trace
2742 settings are restored when the called method returns.
2743 Trace Level
2745 The trace level is the first 4 bits of the trace settings (masked by
2746 "DBIc_TRACE_FLAGS_MASK") and represents trace levels of 1 to 15. Do not
2747 output anything at trace levels less than 3 as they are reserved for
2748 DBI.
2749 For advice on what to output at each level see "Trace Levels" in DBI.
2751 To test for a trace level you can use the "DBIc_TRACE_LEVEL" macro like
2753 this:
2754 if (DBIc_TRACE_LEVEL(imp_xxh) >= 2) {
2756 PerlIO_printf(DBIc_LOGPIO(imp_xxh), "foobar");
2757 }
2758 Also note the use of PerlIO_printf which you should always use for
2760 tracing and never the C "stdio.h" I/O functions.
2761 Trace Flags
2763 Trace flags are used to enable tracing of specific activities within
2764 the DBI and drivers. The DBI defines some trace flags and drivers can
2765 define others. DBI trace flag names begin with a capital letter and
2766 driver specific names begin with a lowercase letter. For a list of DBI
2767 defined trace flags see "Trace Flags" in DBI.
2768 If you want to use private trace flags you'll probably want to be able
2770 to set them by name. Drivers are expected to override the
2771 parse_trace_flag (note the singular) and check if $trace_flag_name is a
2772 driver specific trace flags and, if not, then call the DBIs default
2773 parse_trace_flag(). To do that you'll need to define a
2774 parse_trace_flag() method like this:
2775 sub parse_trace_flag {
2777 my ($h, $name) = @_;
2778 return 0x01000000 if $name eq 'foo';
2779 return 0x02000000 if $name eq 'bar';
2780 return 0x04000000 if $name eq 'baz';
2781 return 0x08000000 if $name eq 'boo';
2782 return 0x10000000 if $name eq 'bop';
2783 return $h->SUPER::parse_trace_flag($name);
2784 }
2785 All private flag names must be lowercase, and all private flags must be
2787 in the top 8 of the 32 bits of "DBIc_TRACE_FLAGS(imp)" i.e.,
2788 0xFF000000.
2789 If you've defined a parse_trace_flag() method in ::db you'll also want
2791 it in ::st, so just alias it in:
2792 *parse_trace_flag = \&DBD::foo:db::parse_trace_flag;
2794 You may want to act on the current 'SQL' trace flag that DBI defines to
2796 output SQL prepared/executed as DBI currently does not do SQL tracing.
2797 Trace Macros
2799 Access to the trace level and trace flags is via a set of macros.
2800 DBIc_TRACE_SETTINGS(imp) returns the trace settings
2802 DBIc_TRACE_LEVEL(imp) returns the trace level
2803 DBIc_TRACE_FLAGS(imp) returns the trace flags
2804 DBIc_TRACE(imp, flags, flaglevel, level)
2805 e.g.,
2807 DBIc_TRACE(imp, 0, 0, 4)
2809 if level >= 4
2810 DBIc_TRACE(imp, DBDtf_FOO, 2, 4)
2812 if tracing DBDtf_FOO & level>=2 or level>=4
2813 DBIc_TRACE(imp, DBDtf_FOO, 2, 0)
2815 as above but never trace just due to level
2816
2818 Study Oraperl.pm (supplied with DBD::Oracle) and Ingperl.pm (supplied
2819 with DBD::Ingres) and the corresponding dbdimp.c files for ideas.
2820 Note that the emulation code sets "$dbh->{CompatMode} = 1;" for each
2822 connection so that the internals of the driver can implement behaviour
2823 compatible with the old interface when dealing with those handles.
2824 Setting emulation perl variables
2826 For example, ingperl has a $sql_rowcount variable. Rather than try to
2827 manually update this in Ingperl.pm it can be done faster in C code. In
2828 "dbd_init()":
2829 sql_rowcount = perl_get_sv("Ingperl::sql_rowcount", GV_ADDMULTI);
2831 In the relevant places do:
2833 if (DBIc_COMPAT(imp_sth)) /* only do this for compatibility mode handles */
2835 sv_setiv(sql_rowcount, the_row_count);
2836
2838 The imp_xyz_t types
2839 Any handle has a corresponding C structure filled with private data.
2840 Some of this data is reserved for use by DBI (except for using the DBIc
2841 macros below), some is for you. See the description of the dbdimp.h
2842 file above for examples. Most functions in dbdimp.c are passed both the
2843 handle "xyz" and a pointer to "imp_xyz". In rare cases, however, you
2844 may use the following macros:
2845 D_imp_dbh(dbh)
2847 Given a function argument dbh, declare a variable imp_dbh and
2848 initialize it with a pointer to the handles private data. Note:
2849 This must be a part of the function header, because it declares a
2850 variable.
2851 D_imp_sth(sth)
2853 Likewise for statement handles.
2854 D_imp_xxx(h)
2856 Given any handle, declare a variable imp_xxx and initialize it with
2857 a pointer to the handles private data. It is safe, for example, to
2858 cast imp_xxx to "imp_dbh_t*", if "DBIc_TYPE(imp_xxx) == DBIt_DB".
2859 (You can also call "sv_derived_from(h, "DBI::db")", but that's much
2860 slower.)
2861 D_imp_dbh_from_sth
2863 Given a imp_sth, declare a variable imp_dbh and initialize it with
2864 a pointer to the parent database handle's implementors structure.
2865 Using DBIc_IMPSET_on
2867 The driver code which initializes a handle should use
2868 "DBIc_IMPSET_on()" as soon as its state is such that the cleanup code
2869 must be called. When this happens is determined by your driver code.
2870 Failure to call this can lead to corruption of data structures.
2872 For example, DBD::Informix maintains a linked list of database handles
2874 in the driver, and within each handle, a linked list of statements.
2875 Once a statement is added to the linked list, it is crucial that it is
2876 cleaned up (removed from the list). When DBIc_IMPSET_on() was being
2877 called too late, it was able to cause all sorts of problems.
2878 Using DBIc_is(), DBIc_has(), DBIc_on() and DBIc_off()
2880 Once upon a long time ago, the only way of handling the internal DBI
2881 boolean flags/attributes was through macros such as:
2882 DBIc_WARN DBIc_WARN_on DBIc_WARN_off
2884 DBIc_COMPAT DBIc_COMPAT_on DBIc_COMPAT_off
2885 Each of these took an imp_xxh pointer as an argument.
2887 Since then, new attributes have been added such as ChopBlanks,
2889 RaiseError and PrintError, and these do not have the full set of
2890 macros. The approved method for handling these is now the four macros:
2891 DBIc_is(imp, flag)
2893 DBIc_has(imp, flag) an alias for DBIc_is
2894 DBIc_on(imp, flag)
2895 DBIc_off(imp, flag)
2896 DBIc_set(imp, flag, on) set if on is true, else clear
2897 Consequently, the "DBIc_XXXXX" family of macros is now mostly
2899 deprecated and new drivers should avoid using them, even though the
2900 older drivers will probably continue to do so for quite a while yet.
2901 However...
2902 There is an important exception to that. The ACTIVE and IMPSET flags
2904 should be set via the "DBIc_ACTIVE_on()" and "DBIc_IMPSET_on()" macros,
2905 and unset via the "DBIc_ACTIVE_off()" and "DBIc_IMPSET_off()" macros.
2906 Using the get_fbav() method
2908 THIS IS CRITICAL for C/XS drivers.
2909 The "$sth->bind_col()" and "$sth->bind_columns()" documented in the DBI
2911 specification do not have to be implemented by the driver writer
2912 because DBI takes care of the details for you.
2913 However, the key to ensuring that bound columns work is to call the
2915 function "DBIc_DBISTATE(imp_xxh)->get_fbav()" in the code which fetches
2916 a row of data.
2917 This returns an "AV", and each element of the "AV" contains the "SV"
2919 which should be set to contain the returned data.
2920 The pure Perl equivalent is the "$sth->_set_fbav($data)" method, as
2922 described in the part on pure Perl drivers.
2923 Casting strings to Perl types based on a SQL type
2925 DBI from 1.611 (and DBIXS_REVISION 13606) defines the
2926 sql_type_cast_svpv method which may be used to cast a string
2927 representation of a value to a more specific Perl type based on a SQL
2928 type. You should consider using this method when processing bound
2929 column data as it provides some support for the TYPE bind_col attribute
2930 which is rarely used in drivers.
2931 int sql_type_cast_svpv(pTHX_ SV *sv, int sql_type, U32 flags, void *v)
2933 "sv" is what you would like cast, "sql_type" is one of the DBI defined
2935 SQL types (e.g., "SQL_INTEGER") and "flags" is a bitmask as follows:
2936 DBIstcf_STRICT
2938 If set this indicates you want an error state returned if the cast
2939 cannot be performed.
2940 DBIstcf_DISCARD_STRING
2942 If set and the pv portion of the "sv" is cast then this will cause
2943 sv's pv to be freed up.
2944 sql_type_cast_svpv returns the following states:
2946 -2 sql_type is not handled - sv not changed
2948 -1 sv is undef, sv not changed
2949 0 sv could not be cast cleanly and DBIstcf_STRICT was specified
2950 1 sv could not be case cleanly and DBIstcf_STRICT was not specified
2951 2 sv was cast ok
2952 The current implementation of sql_type_cast_svpv supports
2954 "SQL_INTEGER", "SQL_DOUBLE" and "SQL_NUMERIC". "SQL_INTEGER" uses
2955 sv_2iv and hence may set IV, UV or NV depending on the number.
2956 "SQL_DOUBLE" uses sv_2nv so may set NV and "SQL_NUMERIC" will set IV or
2957 UV or NV.
2958 DBIstcf_STRICT should be implemented as the StrictlyTyped attribute and
2960 DBIstcf_DISCARD_STRING implemented as the DiscardString attribute to
2961 the bind_col method and both default to off.
2962 See DBD::Oracle for an example of how this is used.
2964
2966 This is definitely an open subject. It can be done, as demonstrated by
2967 the DBD::File driver, but it is not as simple as one might think.
2968 (Note that this topic is different from subclassing the DBI. For an
2970 example of that, see the t/subclass.t file supplied with the DBI.)
2971 The main problem is that the dbh's and sth's that your "connect()" and
2973 "prepare()" methods return are not instances of your DBD::Driver::db or
2974 DBD::Driver::st packages, they are not even derived from it. Instead
2975 they are instances of the DBI::db or DBI::st classes or a derived
2976 subclass. Thus, if you write a method "mymethod()" and do a
2977 $dbh->mymethod()
2979 then the autoloader will search for that method in the package DBI::db.
2981 Of course you can instead to a
2982 $dbh->func('mymethod')
2984 and that will indeed work, even if "mymethod()" is inherited, but not
2986 without additional work. Setting @ISA is not sufficient.
2987 Overwriting methods
2989 The first problem is, that the "connect()" method has no idea of
2990 subclasses. For example, you cannot implement base class and subclass
2991 in the same file: The "install_driver()" method wants to do a
2992 require DBD::Driver;
2994 In particular, your subclass has to be a separate driver, from the view
2996 of DBI, and you cannot share driver handles.
2997 Of course that's not much of a problem. You should even be able to
2999 inherit the base classes "connect()" method. But you cannot simply
3000 overwrite the method, unless you do something like this, quoted from
3001 DBD::CSV:
3002 sub connect ($$;$$$) {
3004 my ($drh, $dbname, $user, $auth, $attr) = @_;
3005 my $this = $drh->DBD::File::dr::connect($dbname, $user, $auth, $attr);
3007 if (!exists($this->{csv_tables})) {
3008 $this->{csv_tables} = {};
3009 }
3010 $this;
3012 }
3013 Note that we cannot do a
3015 $drh->SUPER::connect($dbname, $user, $auth, $attr);
3017 as we would usually do in a an OO environment, because $drh is an
3019 instance of DBI::dr. And note, that the "connect()" method of DBD::File
3020 is able to handle subclass attributes. See the description of Pure Perl
3021 drivers above.
3022 It is essential that you always call superclass method in the above
3024 manner. However, that should do.
3025 Attribute handling
3027 Fortunately the DBI specifications allow a simple, but still performant
3028 way of handling attributes. The idea is based on the convention that
3029 any driver uses a prefix driver_ for its private methods. Thus it's
3030 always clear whether to pass attributes to the super class or not. For
3031 example, consider this "STORE()" method from the DBD::CSV class:
3032 sub STORE {
3034 my ($dbh, $attr, $val) = @_;
3035 if ($attr !~ /^driver_/) {
3036 return $dbh->DBD::File::db::STORE($attr, $val);
3037 }
3038 if ($attr eq 'driver_foo') {
3039 ...
3040 }
3041
3043 Jonathan Leffler <jleffler@us.ibm.com> (previously
3044 <jleffler@informix.com>), Jochen Wiedmann <joe@ispsoft.de>, Steffen
3045 Goeldner <sgoeldner@cpan.org>, and Tim Bunce <dbi-users@perl.org>.
3046perl v5.16.3 2013-04-04 DBI::DBD(3)