1DECLARE(7) PostgreSQL 16.1 Documentation DECLARE(7)
2
6 DECLARE - define a cursor
7
9 DECLARE name [ BINARY ] [ ASENSITIVE | INSENSITIVE ] [ [ NO ] SCROLL ]
10 CURSOR [ { WITH | WITHOUT } HOLD ] FOR query
11
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.
16 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 43.7.
21
23 name
24 The name of the cursor to be created. This must be different from
25 any other active cursor name in the session.
26 BINARY
28 Causes the cursor to return data in binary rather than in text
29 format.
30 ASENSITIVE
32 INSENSITIVE
33 Cursor sensitivity determines whether changes to the data
34 underlying the cursor, done in the same transaction, after the
35 cursor has been declared, are visible in the cursor. INSENSITIVE
36 means they are not visible, ASENSITIVE means the behavior is
37 implementation-dependent. A third behavior, SENSITIVE, meaning that
38 such changes are visible in the cursor, is not available in
39 PostgreSQL. In PostgreSQL, all cursors are insensitive; so these
40 key words have no effect and are only accepted for compatibility
41 with the SQL standard.
42 Specifying INSENSITIVE together with FOR UPDATE or FOR SHARE is an
44 error.
45 SCROLL
47 NO SCROLL
48 SCROLL specifies that the cursor can be used to retrieve rows in a
49 nonsequential fashion (e.g., backward). Depending upon the
50 complexity of the query's execution plan, specifying SCROLL might
51 impose a performance penalty on the query's execution time. NO
52 SCROLL specifies that the cursor cannot be used to retrieve rows in
53 a nonsequential fashion. The default is to allow scrolling in some
54 cases; this is not the same as specifying SCROLL. See Notes below
55 for details.
56 WITH HOLD
58 WITHOUT HOLD
59 WITH HOLD specifies that the cursor can continue to be used after
60 the transaction that created it successfully commits. WITHOUT HOLD
61 specifies that the cursor cannot be used outside of the transaction
62 that created it. If neither WITHOUT HOLD nor WITH HOLD is
63 specified, WITHOUT HOLD is the default.
64 query
66 A SELECT or VALUES command which will provide the rows to be
67 returned by the cursor.
68 The key words ASENSITIVE, BINARY, INSENSITIVE, and SCROLL can appear in
70 any order.
71
73 Normal cursors return data in text format, the same as a SELECT would
74 produce. The BINARY option specifies that the cursor should return data
75 in binary format. This reduces conversion effort for both the server
76 and client, at the cost of more programmer effort to deal with
77 platform-dependent binary data formats. As an example, if a query
78 returns a value of one from an integer column, you would get a string
79 of 1 with a default cursor, whereas with a binary cursor you would get
80 a 4-byte field containing the internal representation of the value (in
81 big-endian byte order).
82 Binary cursors should be used carefully. Many applications, including
84 psql, are not prepared to handle binary cursors and expect data to come
85 back in the text format.
86 Note
88 When the client application uses the “extended query” protocol to
89 issue a FETCH command, the Bind protocol message specifies whether
90 data is to be retrieved in text or binary format. This choice
91 overrides the way that the cursor is defined. The concept of a
92 binary cursor as such is thus obsolete when using extended query
93 protocol — any cursor can be treated as either text or binary.
94 Unless WITH HOLD is specified, the cursor created by this command can
96 only be used within the current transaction. Thus, DECLARE without WITH
97 HOLD is useless outside a transaction block: the cursor would survive
98 only to the completion of the statement. Therefore PostgreSQL reports
99 an error if such a command is used outside a transaction block. Use
100 BEGIN and COMMIT (or ROLLBACK) to define a transaction block.
101 If WITH HOLD is specified and the transaction that created the cursor
103 successfully commits, the cursor can continue to be accessed by
104 subsequent transactions in the same session. (But if the creating
105 transaction is aborted, the cursor is removed.) A cursor created with
106 WITH HOLD is closed when an explicit CLOSE command is issued on it, or
107 the session ends. In the current implementation, the rows represented
108 by a held cursor are copied into a temporary file or memory area so
109 that they remain available for subsequent transactions.
110 WITH HOLD may not be specified when the query includes FOR UPDATE or
112 FOR SHARE.
113 The SCROLL option should be specified when defining a cursor that will
115 be used to fetch backwards. This is required by the SQL standard.
116 However, for compatibility with earlier versions, PostgreSQL will allow
117 backward fetches without SCROLL, if the cursor's query plan is simple
118 enough that no extra overhead is needed to support it. However,
119 application developers are advised not to rely on using backward
120 fetches from a cursor that has not been created with SCROLL. If NO
121 SCROLL is specified, then backward fetches are disallowed in any case.
122 Backward fetches are also disallowed when the query includes FOR UPDATE
124 or FOR SHARE; therefore SCROLL may not be specified in this case.
125 Caution
127 Scrollable cursors may give unexpected results if they invoke any
128 volatile functions (see Section 38.7). When a previously fetched
129 row is re-fetched, the functions might be re-executed, perhaps
130 leading to results different from the first time. It's best to
131 specify NO SCROLL for a query involving volatile functions. If that
132 is not practical, one workaround is to declare the cursor SCROLL
133 WITH HOLD and commit the transaction before reading any rows from
134 it. This will force the entire output of the cursor to be
135 materialized in temporary storage, so that volatile functions are
136 executed exactly once for each row.
137 If the cursor's query includes FOR UPDATE or FOR SHARE, then returned
139 rows are locked at the time they are first fetched, in the same way as
140 for a regular SELECT command with these options. In addition, the
141 returned rows will be the most up-to-date versions.
142 Caution
144 It is generally recommended to use FOR UPDATE if the cursor is
145 intended to be used with UPDATE ... WHERE CURRENT OF or DELETE ...
146 WHERE CURRENT OF. Using FOR UPDATE prevents other sessions from
147 changing the rows between the time they are fetched and the time
148 they are updated. Without FOR UPDATE, a subsequent WHERE CURRENT OF
149 command will have no effect if the row was changed since the cursor
150 was created.
151 Another reason to use FOR UPDATE is that without it, a subsequent
153 WHERE CURRENT OF might fail if the cursor query does not meet the
154 SQL standard's rules for being “simply updatable” (in particular,
155 the cursor must reference just one table and not use grouping or
156 ORDER BY). Cursors that are not simply updatable might work, or
157 might not, depending on plan choice details; so in the worst case,
158 an application might work in testing and then fail in production.
159 If FOR UPDATE is specified, the cursor is guaranteed to be
160 updatable.
161 The main reason not to use FOR UPDATE with WHERE CURRENT OF is if
163 you need the cursor to be scrollable, or to be isolated from
164 concurrent updates (that is, continue to show the old data). If
165 this is a requirement, pay close heed to the caveats shown above.
166 The SQL standard only makes provisions for cursors in embedded SQL. The
168 PostgreSQL server does not implement an OPEN statement for cursors; a
169 cursor is considered to be open when it is declared. However, ECPG, the
170 embedded SQL preprocessor for PostgreSQL, supports the standard SQL
171 cursor conventions, including those involving DECLARE and OPEN
172 statements.
173 The server data structure underlying an open cursor is called a portal.
175 Portal names are exposed in the client protocol: a client can fetch
176 rows directly from an open portal, if it knows the portal name. When
177 creating a cursor with DECLARE, the portal name is the same as the
178 cursor name.
179 You can see all available cursors by querying the pg_cursors system
181 view.
182
184 To declare a cursor:
185 DECLARE liahona CURSOR FOR SELECT * FROM films;
187 See FETCH(7) for more examples of cursor usage.
189
191 The SQL standard allows cursors only in embedded SQL and in modules.
192 PostgreSQL permits cursors to be used interactively.
193 According to the SQL standard, changes made to insensitive cursors by
195 UPDATE ... WHERE CURRENT OF and DELETE ... WHERE CURRENT OF statements
196 are visible in that same cursor. PostgreSQL treats these statements
197 like all other data changing statements in that they are not visible in
198 insensitive cursors.
199 Binary cursors are a PostgreSQL extension.
201
203 CLOSE(7), FETCH(7), MOVE(7)
204PostgreSQL 16.1 2023 DECLARE(7)