1DECLARE(7) PostgreSQL 13.4 Documentation DECLARE(7)
2
6 DECLARE - define a cursor
7
9 DECLARE name [ BINARY ] [ 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(7).
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 42.7.
21
23 name
24 The name of the cursor to be created.
25 BINARY
27 Causes the cursor to return data in binary rather than in text
28 format.
29 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 SCROLL
38 NO SCROLL
39 SCROLL specifies that the cursor can be used to retrieve rows in a
40 nonsequential fashion (e.g., backward). Depending upon the
41 complexity of the query's execution plan, specifying SCROLL might
42 impose a performance penalty on the query's execution time. NO
43 SCROLL specifies that the cursor cannot be used to retrieve rows in
44 a nonsequential fashion. The default is to allow scrolling in some
45 cases; this is not the same as specifying SCROLL. See Notes below
46 for details.
47 WITH HOLD
49 WITHOUT HOLD
50 WITH HOLD specifies that the cursor can continue to be used after
51 the transaction that created it successfully commits. WITHOUT HOLD
52 specifies that the cursor cannot be used outside of the transaction
53 that created it. If neither WITHOUT HOLD nor WITH HOLD is
54 specified, WITHOUT HOLD is the default.
55 query
57 A SELECT(7) or VALUES(7) command which will provide the rows to be
58 returned by the cursor.
59 The key words BINARY, INSENSITIVE, and SCROLL can appear in any order.
61
63 Normal cursors return data in text format, the same as a SELECT would
64 produce. The BINARY option specifies that the cursor should return data
65 in binary format. This reduces conversion effort for both the server
66 and client, at the cost of more programmer effort to deal with
67 platform-dependent binary data formats. As an example, if a query
68 returns a value of one from an integer column, you would get a string
69 of 1 with a default cursor, whereas with a binary cursor you would get
70 a 4-byte field containing the internal representation of the value (in
71 big-endian byte order).
72 Binary cursors should be used carefully. Many applications, including
74 psql, are not prepared to handle binary cursors and expect data to come
75 back in the text format.
76 Note
78 When the client application uses the “extended query” protocol to
79 issue a FETCH command, the Bind protocol message specifies whether
80 data is to be retrieved in text or binary format. This choice
81 overrides the way that the cursor is defined. The concept of a
82 binary cursor as such is thus obsolete when using extended query
83 protocol — any cursor can be treated as either text or binary.
84 Unless WITH HOLD is specified, the cursor created by this command can
86 only be used within the current transaction. Thus, DECLARE without WITH
87 HOLD is useless outside a transaction block: the cursor would survive
88 only to the completion of the statement. Therefore PostgreSQL reports
89 an error if such a command is used outside a transaction block. Use
90 BEGIN(7) and COMMIT(7) (or ROLLBACK(7)) to define a transaction block.
91 If WITH HOLD is specified and the transaction that created the cursor
93 successfully commits, the cursor can continue to be accessed by
94 subsequent transactions in the same session. (But if the creating
95 transaction is aborted, the cursor is removed.) A cursor created with
96 WITH HOLD is closed when an explicit CLOSE command is issued on it, or
97 the session ends. In the current implementation, the rows represented
98 by a held cursor are copied into a temporary file or memory area so
99 that they remain available for subsequent transactions.
100 WITH HOLD may not be specified when the query includes FOR UPDATE or
102 FOR SHARE.
103 The SCROLL option should be specified when defining a cursor that will
105 be used to fetch backwards. This is required by the SQL standard.
106 However, for compatibility with earlier versions, PostgreSQL will allow
107 backward fetches without SCROLL, if the cursor's query plan is simple
108 enough that no extra overhead is needed to support it. However,
109 application developers are advised not to rely on using backward
110 fetches from a cursor that has not been created with SCROLL. If NO
111 SCROLL is specified, then backward fetches are disallowed in any case.
112 Backward fetches are also disallowed when the query includes FOR UPDATE
114 or FOR SHARE; therefore SCROLL may not be specified in this case.
115 Caution
117 Scrollable cursors may give unexpected results if they invoke any
118 volatile functions (see Section 37.7). When a previously fetched
119 row is re-fetched, the functions might be re-executed, perhaps
120 leading to results different from the first time. It's best to
121 specify NO SCROLL for a query involving volatile functions. If that
122 is not practical, one workaround is to declare the cursor SCROLL
123 WITH HOLD and commit the transaction before reading any rows from
124 it. This will force the entire output of the cursor to be
125 materialized in temporary storage, so that volatile functions are
126 executed exactly once for each row.
127 If the cursor's query includes FOR UPDATE or FOR SHARE, then returned
129 rows are locked at the time they are first fetched, in the same way as
130 for a regular SELECT(7) command with these options. In addition, the
131 returned rows will be the most up-to-date versions; therefore these
132 options provide the equivalent of what the SQL standard calls a
133 “sensitive cursor”. (Specifying INSENSITIVE together with FOR UPDATE or
134 FOR SHARE is an error.)
135 Caution
137 It is generally recommended to use FOR UPDATE if the cursor is
138 intended to be used with UPDATE ... WHERE CURRENT OF or DELETE ...
139 WHERE CURRENT OF. Using FOR UPDATE prevents other sessions from
140 changing the rows between the time they are fetched and the time
141 they are updated. Without FOR UPDATE, a subsequent WHERE CURRENT OF
142 command will have no effect if the row was changed since the cursor
143 was created.
144 Another reason to use FOR UPDATE is that without it, a subsequent
146 WHERE CURRENT OF might fail if the cursor query does not meet the
147 SQL standard's rules for being “simply updatable” (in particular,
148 the cursor must reference just one table and not use grouping or
149 ORDER BY). Cursors that are not simply updatable might work, or
150 might not, depending on plan choice details; so in the worst case,
151 an application might work in testing and then fail in production.
152 If FOR UPDATE is specified, the cursor is guaranteed to be
153 updatable.
154 The main reason not to use FOR UPDATE with WHERE CURRENT OF is if
156 you need the cursor to be scrollable, or to be insensitive to the
157 subsequent updates (that is, continue to show the old data). If
158 this is a requirement, pay close heed to the caveats shown above.
159 The SQL standard only makes provisions for cursors in embedded SQL. The
161 PostgreSQL server does not implement an OPEN statement for cursors; a
162 cursor is considered to be open when it is declared. However, ECPG, the
163 embedded SQL preprocessor for PostgreSQL, supports the standard SQL
164 cursor conventions, including those involving DECLARE and OPEN
165 statements.
166 You can see all available cursors by querying the pg_cursors system
168 view.
169
171 To declare a cursor:
172 DECLARE liahona CURSOR FOR SELECT * FROM films;
174 See FETCH(7) for more examples of cursor usage.
176
178 The SQL standard says that it is implementation-dependent whether
179 cursors are sensitive to concurrent updates of the underlying data by
180 default. In PostgreSQL, cursors are insensitive by default, and can be
181 made sensitive by specifying FOR UPDATE. Other products may work
182 differently.
183 The SQL standard allows cursors only in embedded SQL and in modules.
185 PostgreSQL permits cursors to be used interactively.
186 Binary cursors are a PostgreSQL extension.
188
190 CLOSE(7), FETCH(7), MOVE(7)
191PostgreSQL 13.4 2021 DECLARE(7)