1COPY(7) PostgreSQL 12.6 Documentation COPY(7)
2
6 COPY - copy data between a file and a table
7
9 COPY table_name [ ( column_name [, ...] ) ]
10 FROM { 'filename' | PROGRAM 'command' | STDIN }
11 [ [ WITH ] ( option [, ...] ) ]
12 [ WHERE condition ]
13 COPY { table_name [ ( column_name [, ...] ) ] | ( query ) }
15 TO { 'filename' | PROGRAM 'command' | STDOUT }
16 [ [ WITH ] ( option [, ...] ) ]
17 where option can be one of:
19 FORMAT format_name
21 FREEZE [ boolean ]
22 DELIMITER 'delimiter_character'
23 NULL 'null_string'
24 HEADER [ boolean ]
25 QUOTE 'quote_character'
26 ESCAPE 'escape_character'
27 FORCE_QUOTE { ( column_name [, ...] ) | * }
28 FORCE_NOT_NULL ( column_name [, ...] )
29 FORCE_NULL ( column_name [, ...] )
30 ENCODING 'encoding_name'
31
33 COPY moves data between PostgreSQL tables and standard file-system
34 files. COPY TO copies the contents of a table to a file, while COPY
35 FROM copies data from a file to a table (appending the data to whatever
36 is in the table already). COPY TO can also copy the results of a
37 SELECT query.
38 If a column list is specified, COPY TO copies only the data in the
40 specified columns to the file. For COPY FROM, each field in the file is
41 inserted, in order, into the specified column. Table columns not
42 specified in the COPY FROM column list will receive their default
43 values.
44 COPY with a file name instructs the PostgreSQL server to directly read
46 from or write to a file. The file must be accessible by the PostgreSQL
47 user (the user ID the server runs as) and the name must be specified
48 from the viewpoint of the server. When PROGRAM is specified, the server
49 executes the given command and reads from the standard output of the
50 program, or writes to the standard input of the program. The command
51 must be specified from the viewpoint of the server, and be executable
52 by the PostgreSQL user. When STDIN or STDOUT is specified, data is
53 transmitted via the connection between the client and the server.
54
56 table_name
57 The name (optionally schema-qualified) of an existing table.
58 column_name
60 An optional list of columns to be copied. If no column list is
61 specified, all columns of the table except generated columns will
62 be copied.
63 query
65 A SELECT(7), VALUES(7), INSERT(7), UPDATE(7) or DELETE(7) command
66 whose results are to be copied. Note that parentheses are required
67 around the query.
68 For INSERT, UPDATE and DELETE queries a RETURNING clause must be
70 provided, and the target relation must not have a conditional rule,
71 nor an ALSO rule, nor an INSTEAD rule that expands to multiple
72 statements.
73 filename
75 The path name of the input or output file. An input file name can
76 be an absolute or relative path, but an output file name must be an
77 absolute path. Windows users might need to use an E'' string and
78 double any backslashes used in the path name.
79 PROGRAM
81 A command to execute. In COPY FROM, the input is read from standard
82 output of the command, and in COPY TO, the output is written to the
83 standard input of the command.
84 Note that the command is invoked by the shell, so if you need to
86 pass any arguments to shell command that come from an untrusted
87 source, you must be careful to strip or escape any special
88 characters that might have a special meaning for the shell. For
89 security reasons, it is best to use a fixed command string, or at
90 least avoid passing any user input in it.
91 STDIN
93 Specifies that input comes from the client application.
94 STDOUT
96 Specifies that output goes to the client application.
97 boolean
99 Specifies whether the selected option should be turned on or off.
100 You can write TRUE, ON, or 1 to enable the option, and FALSE, OFF,
101 or 0 to disable it. The boolean value can also be omitted, in which
102 case TRUE is assumed.
103 FORMAT
105 Selects the data format to be read or written: text, csv (Comma
106 Separated Values), or binary. The default is text.
107 FREEZE
109 Requests copying the data with rows already frozen, just as they
110 would be after running the VACUUM FREEZE command. This is intended
111 as a performance option for initial data loading. Rows will be
112 frozen only if the table being loaded has been created or truncated
113 in the current subtransaction, there are no cursors open and there
114 are no older snapshots held by this transaction. It is currently
115 not possible to perform a COPY FREEZE on a partitioned table.
116 Note that all other sessions will immediately be able to see the
118 data once it has been successfully loaded. This violates the normal
119 rules of MVCC visibility and users specifying should be aware of
120 the potential problems this might cause.
121 DELIMITER
123 Specifies the character that separates columns within each row
124 (line) of the file. The default is a tab character in text format,
125 a comma in CSV format. This must be a single one-byte character.
126 This option is not allowed when using binary format.
127 NULL
129 Specifies the string that represents a null value. The default is
130 \N (backslash-N) in text format, and an unquoted empty string in
131 CSV format. You might prefer an empty string even in text format
132 for cases where you don't want to distinguish nulls from empty
133 strings. This option is not allowed when using binary format.
134 Note
136 When using COPY FROM, any data item that matches this string
137 will be stored as a null value, so you should make sure that
138 you use the same string as you used with COPY TO.
139 HEADER
141 Specifies that the file contains a header line with the names of
142 each column in the file. On output, the first line contains the
143 column names from the table, and on input, the first line is
144 ignored. This option is allowed only when using CSV format.
145 QUOTE
147 Specifies the quoting character to be used when a data value is
148 quoted. The default is double-quote. This must be a single one-byte
149 character. This option is allowed only when using CSV format.
150 ESCAPE
152 Specifies the character that should appear before a data character
153 that matches the QUOTE value. The default is the same as the QUOTE
154 value (so that the quoting character is doubled if it appears in
155 the data). This must be a single one-byte character. This option is
156 allowed only when using CSV format.
157 FORCE_QUOTE
159 Forces quoting to be used for all non-NULL values in each specified
160 column. NULL output is never quoted. If * is specified, non-NULL
161 values will be quoted in all columns. This option is allowed only
162 in COPY TO, and only when using CSV format.
163 FORCE_NOT_NULL
165 Do not match the specified columns' values against the null string.
166 In the default case where the null string is empty, this means that
167 empty values will be read as zero-length strings rather than nulls,
168 even when they are not quoted. This option is allowed only in COPY
169 FROM, and only when using CSV format.
170 FORCE_NULL
172 Match the specified columns' values against the null string, even
173 if it has been quoted, and if a match is found set the value to
174 NULL. In the default case where the null string is empty, this
175 converts a quoted empty string into NULL. This option is allowed
176 only in COPY FROM, and only when using CSV format.
177 ENCODING
179 Specifies that the file is encoded in the encoding_name. If this
180 option is omitted, the current client encoding is used. See the
181 Notes below for more details.
182 WHERE
184 The optional WHERE clause has the general form
185 WHERE condition
187 where condition is any expression that evaluates to a result of
189 type boolean. Any row that does not satisfy this condition will not
190 be inserted to the table. A row satisfies the condition if it
191 returns true when the actual row values are substituted for any
192 variable references.
193 Currently, subqueries are not allowed in WHERE expressions, and the
195 evaluation does not see any changes made by the COPY itself (this
196 matters when the expression contains calls to VOLATILE functions).
197
199 On successful completion, a COPY command returns a command tag of the
200 form
201 COPY count
203 The count is the number of rows copied.
205 Note
207 psql will print this command tag only if the command was not COPY
208 ... TO STDOUT, or the equivalent psql meta-command \copy ... to
209 stdout. This is to prevent confusing the command tag with the data
210 that was just printed.
211
213 COPY TO can be used only with plain tables, not views, and does not
214 copy rows from child tables or child partitions. For example, COPY
215 table TO copies the same rows as SELECT * FROM ONLY table. The syntax
216 COPY (SELECT * FROM table) TO ... can be used to dump all of the rows
217 in an inheritance hierarchy, partitioned table, or view.
218 COPY FROM can be used with plain, foreign, or partitioned tables or
220 with views that have INSTEAD OF INSERT triggers.
221 You must have select privilege on the table whose values are read by
223 COPY TO, and insert privilege on the table into which values are
224 inserted by COPY FROM. It is sufficient to have column privileges on
225 the column(s) listed in the command.
226 If row-level security is enabled for the table, the relevant SELECT
228 policies will apply to COPY table TO statements. Currently, COPY FROM
229 is not supported for tables with row-level security. Use equivalent
230 INSERT statements instead.
231 Files named in a COPY command are read or written directly by the
233 server, not by the client application. Therefore, they must reside on
234 or be accessible to the database server machine, not the client. They
235 must be accessible to and readable or writable by the PostgreSQL user
236 (the user ID the server runs as), not the client. Similarly, the
237 command specified with PROGRAM is executed directly by the server, not
238 by the client application, must be executable by the PostgreSQL user.
239 COPY naming a file or command is only allowed to database superusers or
240 users who are granted one of the default roles pg_read_server_files,
241 pg_write_server_files, or pg_execute_server_program, since it allows
242 reading or writing any file or running a program that the server has
243 privileges to access.
244 Do not confuse COPY with the psql instruction \copy. \copy invokes
246 COPY FROM STDIN or COPY TO STDOUT, and then fetches/stores the data in
247 a file accessible to the psql client. Thus, file accessibility and
248 access rights depend on the client rather than the server when \copy is
249 used.
250 It is recommended that the file name used in COPY always be specified
252 as an absolute path. This is enforced by the server in the case of COPY
253 TO, but for COPY FROM you do have the option of reading from a file
254 specified by a relative path. The path will be interpreted relative to
255 the working directory of the server process (normally the cluster's
256 data directory), not the client's working directory.
257 Executing a command with PROGRAM might be restricted by the operating
259 system's access control mechanisms, such as SELinux.
260 COPY FROM will invoke any triggers and check constraints on the
262 destination table. However, it will not invoke rules.
263 For identity columns, the COPY FROM command will always write the
265 column values provided in the input data, like the INSERT option
266 OVERRIDING SYSTEM VALUE.
267 COPY input and output is affected by DateStyle. To ensure portability
269 to other PostgreSQL installations that might use non-default DateStyle
270 settings, DateStyle should be set to ISO before using COPY TO. It is
271 also a good idea to avoid dumping data with IntervalStyle set to
272 sql_standard, because negative interval values might be misinterpreted
273 by a server that has a different setting for IntervalStyle.
274 Input data is interpreted according to ENCODING option or the current
276 client encoding, and output data is encoded in ENCODING or the current
277 client encoding, even if the data does not pass through the client but
278 is read from or written to a file directly by the server.
279 COPY stops operation at the first error. This should not lead to
281 problems in the event of a COPY TO, but the target table will already
282 have received earlier rows in a COPY FROM. These rows will not be
283 visible or accessible, but they still occupy disk space. This might
284 amount to a considerable amount of wasted disk space if the failure
285 happened well into a large copy operation. You might wish to invoke
286 VACUUM to recover the wasted space.
287 FORCE_NULL and FORCE_NOT_NULL can be used simultaneously on the same
289 column. This results in converting quoted null strings to null values
290 and unquoted null strings to empty strings.
291
293 Text Format
294 When the text format is used, the data read or written is a text file
295 with one line per table row. Columns in a row are separated by the
296 delimiter character. The column values themselves are strings generated
297 by the output function, or acceptable to the input function, of each
298 attribute's data type. The specified null string is used in place of
299 columns that are null. COPY FROM will raise an error if any line of
300 the input file contains more or fewer columns than are expected.
301 End of data can be represented by a single line containing just
303 backslash-period (\.). An end-of-data marker is not necessary when
304 reading from a file, since the end of file serves perfectly well; it is
305 needed only when copying data to or from client applications using
306 pre-3.0 client protocol.
307 Backslash characters (\) can be used in the COPY data to quote data
309 characters that might otherwise be taken as row or column delimiters.
310 In particular, the following characters must be preceded by a backslash
311 if they appear as part of a column value: backslash itself, newline,
312 carriage return, and the current delimiter character.
313 The specified null string is sent by COPY TO without adding any
315 backslashes; conversely, COPY FROM matches the input against the null
316 string before removing backslashes. Therefore, a null string such as \N
317 cannot be confused with the actual data value \N (which would be
318 represented as \\N).
319 The following special backslash sequences are recognized by COPY FROM:
321 ┌─────────┬────────────────────────────┐
323 │Sequence │ Represents │
324 ├─────────┼────────────────────────────┤
325 │\b │ Backspace (ASCII 8) │
326 ├─────────┼────────────────────────────┤
327 │\f │ Form feed (ASCII 12) │
328 ├─────────┼────────────────────────────┤
329 │\n │ Newline (ASCII 10) │
330 ├─────────┼────────────────────────────┤
331 │\r │ Carriage return (ASCII 13) │
332 ├─────────┼────────────────────────────┤
333 │\t │ Tab (ASCII 9) │
334 ├─────────┼────────────────────────────┤
335 │\v │ Vertical tab (ASCII 11) │
336 ├─────────┼────────────────────────────┤
337 │\digits │ Backslash followed by one │
338 │ │ to three octal digits │
339 │ │ specifies │
340 │ │ the character with │
341 │ │ that numeric code │
342 ├─────────┼────────────────────────────┤
343 │\xdigits │ Backslash x followed by │
344 │ │ one or two hex digits │
345 │ │ specifies │
346 │ │ the character with │
347 │ │ that numeric code │
348 └─────────┴────────────────────────────┘
349 Presently, COPY TO will never emit an octal or hex-digits backslash
350 sequence, but it does use the other sequences listed above for those
351 control characters.
352 Any other backslashed character that is not mentioned in the above
354 table will be taken to represent itself. However, beware of adding
355 backslashes unnecessarily, since that might accidentally produce a
356 string matching the end-of-data marker (\.) or the null string (\N by
357 default). These strings will be recognized before any other backslash
358 processing is done.
359 It is strongly recommended that applications generating COPY data
361 convert data newlines and carriage returns to the \n and \r sequences
362 respectively. At present it is possible to represent a data carriage
363 return by a backslash and carriage return, and to represent a data
364 newline by a backslash and newline. However, these representations
365 might not be accepted in future releases. They are also highly
366 vulnerable to corruption if the COPY file is transferred across
367 different machines (for example, from Unix to Windows or vice versa).
368 COPY TO will terminate each row with a Unix-style newline (“\n”).
370 Servers running on Microsoft Windows instead output carriage
371 return/newline (“\r\n”), but only for COPY to a server file; for
372 consistency across platforms, COPY TO STDOUT always sends “\n”
373 regardless of server platform. COPY FROM can handle lines ending with
374 newlines, carriage returns, or carriage return/newlines. To reduce the
375 risk of error due to un-backslashed newlines or carriage returns that
376 were meant as data, COPY FROM will complain if the line endings in the
377 input are not all alike.
378 CSV Format
380 This format option is used for importing and exporting the Comma
381 Separated Value (CSV) file format used by many other programs, such as
382 spreadsheets. Instead of the escaping rules used by PostgreSQL's
383 standard text format, it produces and recognizes the common CSV
384 escaping mechanism.
385 The values in each record are separated by the DELIMITER character. If
387 the value contains the delimiter character, the QUOTE character, the
388 NULL string, a carriage return, or line feed character, then the whole
389 value is prefixed and suffixed by the QUOTE character, and any
390 occurrence within the value of a QUOTE character or the ESCAPE
391 character is preceded by the escape character. You can also use
392 FORCE_QUOTE to force quotes when outputting non-NULL values in specific
393 columns.
394 The CSV format has no standard way to distinguish a NULL value from an
396 empty string. PostgreSQL's COPY handles this by quoting. A NULL is
397 output as the NULL parameter string and is not quoted, while a non-NULL
398 value matching the NULL parameter string is quoted. For example, with
399 the default settings, a NULL is written as an unquoted empty string,
400 while an empty string data value is written with double quotes ("").
401 Reading values follows similar rules. You can use FORCE_NOT_NULL to
402 prevent NULL input comparisons for specific columns. You can also use
403 FORCE_NULL to convert quoted null string data values to NULL.
404 Because backslash is not a special character in the CSV format, \., the
406 end-of-data marker, could also appear as a data value. To avoid any
407 misinterpretation, a \. data value appearing as a lone entry on a line
408 is automatically quoted on output, and on input, if quoted, is not
409 interpreted as the end-of-data marker. If you are loading a file
410 created by another application that has a single unquoted column and
411 might have a value of \., you might need to quote that value in the
412 input file.
413 Note
415 In CSV format, all characters are significant. A quoted value
416 surrounded by white space, or any characters other than DELIMITER,
417 will include those characters. This can cause errors if you import
418 data from a system that pads CSV lines with white space out to some
419 fixed width. If such a situation arises you might need to
420 preprocess the CSV file to remove the trailing white space, before
421 importing the data into PostgreSQL.
422 Note
424 CSV format will both recognize and produce CSV files with quoted
425 values containing embedded carriage returns and line feeds. Thus
426 the files are not strictly one line per table row like text-format
427 files.
428 Note
430 Many programs produce strange and occasionally perverse CSV files,
431 so the file format is more a convention than a standard. Thus you
432 might encounter some files that cannot be imported using this
433 mechanism, and COPY might produce files that other programs cannot
434 process.
435 Binary Format
437 The binary format option causes all data to be stored/read as binary
438 format rather than as text. It is somewhat faster than the text and CSV
439 formats, but a binary-format file is less portable across machine
440 architectures and PostgreSQL versions. Also, the binary format is very
441 data type specific; for example it will not work to output binary data
442 from a smallint column and read it into an integer column, even though
443 that would work fine in text format.
444 The binary file format consists of a file header, zero or more tuples
446 containing the row data, and a file trailer. Headers and data are in
447 network byte order.
448 Note
450 PostgreSQL releases before 7.4 used a different binary file format.
451 File Header
453 The file header consists of 15 bytes of fixed fields, followed by a
454 variable-length header extension area. The fixed fields are:
455 Signature
457 11-byte sequence PGCOPY\n\377\r\n\0 — note that the zero byte
458 is a required part of the signature. (The signature is designed
459 to allow easy identification of files that have been munged by
460 a non-8-bit-clean transfer. This signature will be changed by
461 end-of-line-translation filters, dropped zero bytes, dropped
462 high bits, or parity changes.)
463 Flags field
465 32-bit integer bit mask to denote important aspects of the file
466 format. Bits are numbered from 0 (LSB) to 31 (MSB). Note that
467 this field is stored in network byte order (most significant
468 byte first), as are all the integer fields used in the file
469 format. Bits 16-31 are reserved to denote critical file format
470 issues; a reader should abort if it finds an unexpected bit set
471 in this range. Bits 0-15 are reserved to signal
472 backwards-compatible format issues; a reader should simply
473 ignore any unexpected bits set in this range. Currently only
474 one flag bit is defined, and the rest must be zero:
475 Bit 16
477 If 1, OIDs are included in the data; if 0, not. Oid system
478 columns are not supported in PostgreSQL anymore, but the
479 format still contains the indicator.
480 Header extension area length
482 32-bit integer, length in bytes of remainder of header, not
483 including self. Currently, this is zero, and the first tuple
484 follows immediately. Future changes to the format might allow
485 additional data to be present in the header. A reader should
486 silently skip over any header extension data it does not know
487 what to do with.
488 The header extension area is envisioned to contain a sequence of
490 self-identifying chunks. The flags field is not intended to tell
491 readers what is in the extension area. Specific design of header
492 extension contents is left for a later release.
493 This design allows for both backwards-compatible header additions
495 (add header extension chunks, or set low-order flag bits) and
496 non-backwards-compatible changes (set high-order flag bits to
497 signal such changes, and add supporting data to the extension area
498 if needed).
499 Tuples
501 Each tuple begins with a 16-bit integer count of the number of
502 fields in the tuple. (Presently, all tuples in a table will have
503 the same count, but that might not always be true.) Then, repeated
504 for each field in the tuple, there is a 32-bit length word followed
505 by that many bytes of field data. (The length word does not include
506 itself, and can be zero.) As a special case, -1 indicates a NULL
507 field value. No value bytes follow in the NULL case.
508 There is no alignment padding or any other extra data between
510 fields.
511 Presently, all data values in a binary-format file are assumed to
513 be in binary format (format code one). It is anticipated that a
514 future extension might add a header field that allows per-column
515 format codes to be specified.
516 To determine the appropriate binary format for the actual tuple
518 data you should consult the PostgreSQL source, in particular the
519 *send and *recv functions for each column's data type (typically
520 these functions are found in the src/backend/utils/adt/ directory
521 of the source distribution).
522 If OIDs are included in the file, the OID field immediately follows
524 the field-count word. It is a normal field except that it's not
525 included in the field-count. Note that oid system columns are not
526 supported in current versions of PostgreSQL.
527 File Trailer
529 The file trailer consists of a 16-bit integer word containing -1.
530 This is easily distinguished from a tuple's field-count word.
531 A reader should report an error if a field-count word is neither -1
533 nor the expected number of columns. This provides an extra check
534 against somehow getting out of sync with the data.
535
537 The following example copies a table to the client using the vertical
538 bar (|) as the field delimiter:
539 COPY country TO STDOUT (DELIMITER '|');
541 To copy data from a file into the country table:
543 COPY country FROM '/usr1/proj/bray/sql/country_data';
545 To copy into a file just the countries whose names start with 'A':
547 COPY (SELECT * FROM country WHERE country_name LIKE 'A%') TO '/usr1/proj/bray/sql/a_list_countries.copy';
549 To copy into a compressed file, you can pipe the output through an
551 external compression program:
552 COPY country TO PROGRAM 'gzip > /usr1/proj/bray/sql/country_data.gz';
554 Here is a sample of data suitable for copying into a table from STDIN:
556 AF AFGHANISTAN
558 AL ALBANIA
559 DZ ALGERIA
560 ZM ZAMBIA
561 ZW ZIMBABWE
562 Note that the white space on each line is actually a tab character.
564 The following is the same data, output in binary format. The data is
566 shown after filtering through the Unix utility od -c. The table has
567 three columns; the first has type char(2), the second has type text,
568 and the third has type integer. All the rows have a null value in the
569 third column.
570 0000000 P G C O P Y \n 377 \r \n \0 \0 \0 \0 \0 \0
572 0000020 \0 \0 \0 \0 003 \0 \0 \0 002 A F \0 \0 \0 013 A
573 0000040 F G H A N I S T A N 377 377 377 377 \0 003
574 0000060 \0 \0 \0 002 A L \0 \0 \0 007 A L B A N I
575 0000100 A 377 377 377 377 \0 003 \0 \0 \0 002 D Z \0 \0 \0
576 0000120 007 A L G E R I A 377 377 377 377 \0 003 \0 \0
577 0000140 \0 002 Z M \0 \0 \0 006 Z A M B I A 377 377
578 0000160 377 377 \0 003 \0 \0 \0 002 Z W \0 \0 \0 \b Z I
579 0000200 M B A B W E 377 377 377 377 377 377
580
582 There is no COPY statement in the SQL standard.
583 The following syntax was used before PostgreSQL version 9.0 and is
585 still supported:
586 COPY table_name [ ( column_name [, ...] ) ]
588 FROM { 'filename' | STDIN }
589 [ [ WITH ]
590 [ BINARY ]
591 [ DELIMITER [ AS ] 'delimiter_character' ]
592 [ NULL [ AS ] 'null_string' ]
593 [ CSV [ HEADER ]
594 [ QUOTE [ AS ] 'quote_character' ]
595 [ ESCAPE [ AS ] 'escape_character' ]
596 [ FORCE NOT NULL column_name [, ...] ] ] ]
597 COPY { table_name [ ( column_name [, ...] ) ] | ( query ) }
599 TO { 'filename' | STDOUT }
600 [ [ WITH ]
601 [ BINARY ]
602 [ DELIMITER [ AS ] 'delimiter_character' ]
603 [ NULL [ AS ] 'null_string' ]
604 [ CSV [ HEADER ]
605 [ QUOTE [ AS ] 'quote_character' ]
606 [ ESCAPE [ AS ] 'escape_character' ]
607 [ FORCE QUOTE { column_name [, ...] | * } ] ] ]
608 Note that in this syntax, BINARY and CSV are treated as independent
610 keywords, not as arguments of a FORMAT option.
611 The following syntax was used before PostgreSQL version 7.3 and is
613 still supported:
614 COPY [ BINARY ] table_name
616 FROM { 'filename' | STDIN }
617 [ [USING] DELIMITERS 'delimiter_character' ]
618 [ WITH NULL AS 'null_string' ]
619 COPY [ BINARY ] table_name
621 TO { 'filename' | STDOUT }
622 [ [USING] DELIMITERS 'delimiter_character' ]
623 [ WITH NULL AS 'null_string' ]
624PostgreSQL 12.6 2021 COPY(7)