1DECLARE(7)              PostgreSQL 9.2.24 Documentation             DECLARE(7)
2
3
4

NAME

6       DECLARE - define a cursor
7

SYNOPSIS

9       DECLARE name [ BINARY ] [ INSENSITIVE ] [ [ NO ] SCROLL ]
10           CURSOR [ { WITH | WITHOUT } HOLD ] FOR query
11

DESCRIPTION

13       DECLARE allows a user to create cursors, which can be used to retrieve
14       a small number of rows at a time out of a larger query. After the
15       cursor is created, rows are fetched from it using FETCH(7).
16
17           Note
18           This page describes usage of cursors at the SQL command level. If
19           you are trying to use cursors inside a PL/pgSQL function, the rules
20           are different — see Section 39.7, “Cursors”, in the documentation.
21

PARAMETERS

23       name
24           The name of the cursor to be created.
25
26       BINARY
27           Causes the cursor to return data in binary rather than in text
28           format.
29
30       INSENSITIVE
31           Indicates that data retrieved from the cursor should be unaffected
32           by updates to the table(s) underlying the cursor that occur after
33           the cursor is created. In PostgreSQL, this is the default behavior;
34           so this key word has no effect and is only accepted for
35           compatibility with the SQL standard.
36
37       SCROLL, NO SCROLL
38           SCROLL specifies that the cursor can be used to retrieve rows in a
39           nonsequential fashion (e.g., backward). Depending upon the
40           complexity of the query's execution plan, specifying SCROLL might
41           impose a performance penalty on the query's execution time.  NO
42           SCROLL specifies that the cursor cannot be used to retrieve rows in
43           a nonsequential fashion. The default is to allow scrolling in some
44           cases; this is not the same as specifying SCROLL. See NOTES for
45           details.
46
47       WITH HOLD, WITHOUT HOLD
48           WITH HOLD specifies that the cursor can continue to be used after
49           the transaction that created it successfully commits.  WITHOUT HOLD
50           specifies that the cursor cannot be used outside of the transaction
51           that created it. If neither WITHOUT HOLD nor WITH HOLD is
52           specified, WITHOUT HOLD is the default.
53
54       query
55           A SELECT(7) or VALUES(7) command which will provide the rows to be
56           returned by the cursor.
57
58       The key words BINARY, INSENSITIVE, and SCROLL can appear in any order.
59

NOTES

61       Normal cursors return data in text format, the same as a SELECT would
62       produce. The BINARY option specifies that the cursor should return data
63       in binary format. This reduces conversion effort for both the server
64       and client, at the cost of more programmer effort to deal with
65       platform-dependent binary data formats. As an example, if a query
66       returns a value of one from an integer column, you would get a string
67       of 1 with a default cursor, whereas with a binary cursor you would get
68       a 4-byte field containing the internal representation of the value (in
69       big-endian byte order).
70
71       Binary cursors should be used carefully. Many applications, including
72       psql, are not prepared to handle binary cursors and expect data to come
73       back in the text format.
74
75           Note
76           When the client application uses the “extended query” protocol to
77           issue a FETCH command, the Bind protocol message specifies whether
78           data is to be retrieved in text or binary format. This choice
79           overrides the way that the cursor is defined. The concept of a
80           binary cursor as such is thus obsolete when using extended query
81           protocol — any cursor can be treated as either text or binary.
82
83       Unless WITH HOLD is specified, the cursor created by this command can
84       only be used within the current transaction. Thus, DECLARE without WITH
85       HOLD is useless outside a transaction block: the cursor would survive
86       only to the completion of the statement. Therefore PostgreSQL reports
87       an error if such a command is used outside a transaction block. Use
88       BEGIN(7) and COMMIT(7) (or ROLLBACK(7)) to define a transaction block.
89
90       If WITH HOLD is specified and the transaction that created the cursor
91       successfully commits, the cursor can continue to be accessed by
92       subsequent transactions in the same session. (But if the creating
93       transaction is aborted, the cursor is removed.) A cursor created with
94       WITH HOLD is closed when an explicit CLOSE command is issued on it, or
95       the session ends. In the current implementation, the rows represented
96       by a held cursor are copied into a temporary file or memory area so
97       that they remain available for subsequent transactions.
98
99       WITH HOLD may not be specified when the query includes FOR UPDATE or
100       FOR SHARE.
101
102       The SCROLL option should be specified when defining a cursor that will
103       be used to fetch backwards. This is required by the SQL standard.
104       However, for compatibility with earlier versions, PostgreSQL will allow
105       backward fetches without SCROLL, if the cursor's query plan is simple
106       enough that no extra overhead is needed to support it. However,
107       application developers are advised not to rely on using backward
108       fetches from a cursor that has not been created with SCROLL. If NO
109       SCROLL is specified, then backward fetches are disallowed in any case.
110
111       Backward fetches are also disallowed when the query includes FOR UPDATE
112       or FOR SHARE; therefore SCROLL may not be specified in this case.
113
114           Caution
115           Scrollable and WITH HOLD cursors may give unexpected results if
116           they invoke any volatile functions (see Section 35.6, “Function
117           Volatility Categories”, in the documentation). When a previously
118           fetched row is re-fetched, the functions might be re-executed,
119           perhaps leading to results different from the first time. One
120           workaround for such cases is to declare the cursor WITH HOLD and
121           commit the transaction before reading any rows from it. This will
122           force the entire output of the cursor to be materialized in
123           temporary storage, so that volatile functions are executed exactly
124           once for each row.
125
126       If the cursor's query includes FOR UPDATE or FOR SHARE, then returned
127       rows are locked at the time they are first fetched, in the same way as
128       for a regular SELECT(7) command with these options. In addition, the
129       returned rows will be the most up-to-date versions; therefore these
130       options provide the equivalent of what the SQL standard calls a
131       “sensitive cursor”. (Specifying INSENSITIVE together with FOR UPDATE or
132       FOR SHARE is an error.)
133
134           Caution
135           It is generally recommended to use FOR UPDATE if the cursor is
136           intended to be used with UPDATE ... WHERE CURRENT OF or DELETE ...
137           WHERE CURRENT OF. Using FOR UPDATE prevents other sessions from
138           changing the rows between the time they are fetched and the time
139           they are updated. Without FOR UPDATE, a subsequent WHERE CURRENT OF
140           command will have no effect if the row was changed since the cursor
141           was created.
142
143           Another reason to use FOR UPDATE is that without it, a subsequent
144           WHERE CURRENT OF might fail if the cursor query does not meet the
145           SQL standard's rules for being “simply updatable” (in particular,
146           the cursor must reference just one table and not use grouping or
147           ORDER BY). Cursors that are not simply updatable might work, or
148           might not, depending on plan choice details; so in the worst case,
149           an application might work in testing and then fail in production.
150
151           The main reason not to use FOR UPDATE with WHERE CURRENT OF is if
152           you need the cursor to be scrollable, or to be insensitive to the
153           subsequent updates (that is, continue to show the old data). If
154           this is a requirement, pay close heed to the caveats shown above.
155
156       The SQL standard only makes provisions for cursors in embedded SQL. The
157       PostgreSQL server does not implement an OPEN statement for cursors; a
158       cursor is considered to be open when it is declared. However, ECPG, the
159       embedded SQL preprocessor for PostgreSQL, supports the standard SQL
160       cursor conventions, including those involving DECLARE and OPEN
161       statements.
162
163       You can see all available cursors by querying the pg_cursors system
164       view.
165

EXAMPLES

167       To declare a cursor:
168
169           DECLARE liahona CURSOR FOR SELECT * FROM films;
170
171       See FETCH(7) for more examples of cursor usage.
172

COMPATIBILITY

174       The SQL standard says that it is implementation-dependent whether
175       cursors are sensitive to concurrent updates of the underlying data by
176       default. In PostgreSQL, cursors are insensitive by default, and can be
177       made sensitive by specifying FOR UPDATE. Other products may work
178       differently.
179
180       The SQL standard allows cursors only in embedded SQL and in modules.
181       PostgreSQL permits cursors to be used interactively.
182
183       Binary cursors are a PostgreSQL extension.
184

SEE ALSO

186       CLOSE(7), FETCH(7), MOVE(7)
187
188
189
190PostgreSQL 9.2.24                 2017-11-06                        DECLARE(7)
Impressum