| Current File : //usr/share/doc/postgresql-9.2.24/html/sql-select.html |
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML
><HEAD
><TITLE
>SELECT</TITLE
><META
NAME="GENERATOR"
CONTENT="Modular DocBook HTML Stylesheet Version 1.79"><LINK
REV="MADE"
HREF="mailto:pgsql-docs@postgresql.org"><LINK
REL="HOME"
TITLE="PostgreSQL 9.2.24 Documentation"
HREF="index.html"><LINK
REL="UP"
TITLE="SQL Commands"
HREF="sql-commands.html"><LINK
REL="PREVIOUS"
TITLE="SECURITY LABEL"
HREF="sql-security-label.html"><LINK
REL="NEXT"
TITLE="SELECT INTO"
HREF="sql-selectinto.html"><LINK
REL="STYLESHEET"
TYPE="text/css"
HREF="stylesheet.css"><META
HTTP-EQUIV="Content-Type"
CONTENT="text/html; charset=ISO-8859-1"><META
NAME="creation"
CONTENT="2017-11-06T22:43:11"></HEAD
><BODY
CLASS="REFENTRY"
><DIV
CLASS="NAVHEADER"
><TABLE
SUMMARY="Header navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TH
COLSPAN="5"
ALIGN="center"
VALIGN="bottom"
><A
HREF="index.html"
>PostgreSQL 9.2.24 Documentation</A
></TH
></TR
><TR
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
TITLE="SECURITY LABEL"
HREF="sql-security-label.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
HREF="sql-commands.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="60%"
ALIGN="center"
VALIGN="bottom"
></TD
><TD
WIDTH="20%"
ALIGN="right"
VALIGN="top"
><A
TITLE="SELECT INTO"
HREF="sql-selectinto.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><H1
><A
NAME="SQL-SELECT"
></A
>SELECT</H1
><DIV
CLASS="REFNAMEDIV"
><A
NAME="AEN77583"
></A
><H2
>Name</H2
>SELECT, TABLE, WITH -- retrieve rows from a table or view</DIV
><DIV
CLASS="REFSYNOPSISDIV"
><A
NAME="AEN77595"
></A
><H2
>Synopsis</H2
><PRE
CLASS="SYNOPSIS"
>[ WITH [ RECURSIVE ] <TT
CLASS="REPLACEABLE"
><I
>with_query</I
></TT
> [, ...] ]
SELECT [ ALL | DISTINCT [ ON ( <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [, ...] ) ] ]
* | <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [ [ AS ] <TT
CLASS="REPLACEABLE"
><I
>output_name</I
></TT
> ] [, ...]
[ FROM <TT
CLASS="REPLACEABLE"
><I
>from_item</I
></TT
> [, ...] ]
[ WHERE <TT
CLASS="REPLACEABLE"
><I
>condition</I
></TT
> ]
[ GROUP BY <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [, ...] ]
[ HAVING <TT
CLASS="REPLACEABLE"
><I
>condition</I
></TT
> [, ...] ]
[ WINDOW <TT
CLASS="REPLACEABLE"
><I
>window_name</I
></TT
> AS ( <TT
CLASS="REPLACEABLE"
><I
>window_definition</I
></TT
> ) [, ...] ]
[ { UNION | INTERSECT | EXCEPT } [ ALL | DISTINCT ] <TT
CLASS="REPLACEABLE"
><I
>select</I
></TT
> ]
[ ORDER BY <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [ ASC | DESC | USING <TT
CLASS="REPLACEABLE"
><I
>operator</I
></TT
> ] [ NULLS { FIRST | LAST } ] [, ...] ]
[ LIMIT { <TT
CLASS="REPLACEABLE"
><I
>count</I
></TT
> | ALL } ]
[ OFFSET <TT
CLASS="REPLACEABLE"
><I
>start</I
></TT
> [ ROW | ROWS ] ]
[ FETCH { FIRST | NEXT } [ <TT
CLASS="REPLACEABLE"
><I
>count</I
></TT
> ] { ROW | ROWS } ONLY ]
[ FOR { UPDATE | SHARE } [ OF <TT
CLASS="REPLACEABLE"
><I
>table_name</I
></TT
> [, ...] ] [ NOWAIT ] [...] ]
<SPAN
CLASS="phrase"
><SPAN
CLASS="PHRASE"
>where <TT
CLASS="REPLACEABLE"
><I
>from_item</I
></TT
> can be one of:</SPAN
></SPAN
>
[ ONLY ] <TT
CLASS="REPLACEABLE"
><I
>table_name</I
></TT
> [ * ] [ [ AS ] <TT
CLASS="REPLACEABLE"
><I
>alias</I
></TT
> [ ( <TT
CLASS="REPLACEABLE"
><I
>column_alias</I
></TT
> [, ...] ) ] ]
( <TT
CLASS="REPLACEABLE"
><I
>select</I
></TT
> ) [ AS ] <TT
CLASS="REPLACEABLE"
><I
>alias</I
></TT
> [ ( <TT
CLASS="REPLACEABLE"
><I
>column_alias</I
></TT
> [, ...] ) ]
<TT
CLASS="REPLACEABLE"
><I
>with_query_name</I
></TT
> [ [ AS ] <TT
CLASS="REPLACEABLE"
><I
>alias</I
></TT
> [ ( <TT
CLASS="REPLACEABLE"
><I
>column_alias</I
></TT
> [, ...] ) ] ]
<TT
CLASS="REPLACEABLE"
><I
>function_name</I
></TT
> ( [ <TT
CLASS="REPLACEABLE"
><I
>argument</I
></TT
> [, ...] ] ) [ AS ] <TT
CLASS="REPLACEABLE"
><I
>alias</I
></TT
> [ ( <TT
CLASS="REPLACEABLE"
><I
>column_alias</I
></TT
> [, ...] | <TT
CLASS="REPLACEABLE"
><I
>column_definition</I
></TT
> [, ...] ) ]
<TT
CLASS="REPLACEABLE"
><I
>function_name</I
></TT
> ( [ <TT
CLASS="REPLACEABLE"
><I
>argument</I
></TT
> [, ...] ] ) AS ( <TT
CLASS="REPLACEABLE"
><I
>column_definition</I
></TT
> [, ...] )
<TT
CLASS="REPLACEABLE"
><I
>from_item</I
></TT
> [ NATURAL ] <TT
CLASS="REPLACEABLE"
><I
>join_type</I
></TT
> <TT
CLASS="REPLACEABLE"
><I
>from_item</I
></TT
> [ ON <TT
CLASS="REPLACEABLE"
><I
>join_condition</I
></TT
> | USING ( <TT
CLASS="REPLACEABLE"
><I
>join_column</I
></TT
> [, ...] ) ]
<SPAN
CLASS="phrase"
><SPAN
CLASS="PHRASE"
>and <TT
CLASS="REPLACEABLE"
><I
>with_query</I
></TT
> is:</SPAN
></SPAN
>
<TT
CLASS="REPLACEABLE"
><I
>with_query_name</I
></TT
> [ ( <TT
CLASS="REPLACEABLE"
><I
>column_name</I
></TT
> [, ...] ) ] AS ( <TT
CLASS="REPLACEABLE"
><I
>select</I
></TT
> | <TT
CLASS="REPLACEABLE"
><I
>values</I
></TT
> | <TT
CLASS="REPLACEABLE"
><I
>insert</I
></TT
> | <TT
CLASS="REPLACEABLE"
><I
>update</I
></TT
> | <TT
CLASS="REPLACEABLE"
><I
>delete</I
></TT
> )
TABLE [ ONLY ] <TT
CLASS="REPLACEABLE"
><I
>table_name</I
></TT
> [ * ]</PRE
></DIV
><DIV
CLASS="REFSECT1"
><A
NAME="AEN77648"
></A
><H2
>Description</H2
><P
> <TT
CLASS="COMMAND"
>SELECT</TT
> retrieves rows from zero or more tables.
The general processing of <TT
CLASS="COMMAND"
>SELECT</TT
> is as follows:
<P
></P
></P><OL
TYPE="1"
><LI
><P
> All queries in the <TT
CLASS="LITERAL"
>WITH</TT
> list are computed.
These effectively serve as temporary tables that can be referenced
in the <TT
CLASS="LITERAL"
>FROM</TT
> list. A <TT
CLASS="LITERAL"
>WITH</TT
> query
that is referenced more than once in <TT
CLASS="LITERAL"
>FROM</TT
> is
computed only once.
(See <A
HREF="sql-select.html#SQL-WITH"
><I
><I
>WITH</I
> Clause</I
></A
> below.)
</P
></LI
><LI
><P
> All elements in the <TT
CLASS="LITERAL"
>FROM</TT
> list are computed.
(Each element in the <TT
CLASS="LITERAL"
>FROM</TT
> list is a real or
virtual table.) If more than one element is specified in the
<TT
CLASS="LITERAL"
>FROM</TT
> list, they are cross-joined together.
(See <A
HREF="sql-select.html#SQL-FROM"
><I
><I
>FROM</I
> Clause</I
></A
> below.)
</P
></LI
><LI
><P
> If the <TT
CLASS="LITERAL"
>WHERE</TT
> clause is specified, all rows
that do not satisfy the condition are eliminated from the
output. (See <A
HREF="sql-select.html#SQL-WHERE"
><I
><I
>WHERE</I
> Clause</I
></A
> below.)
</P
></LI
><LI
><P
> If the <TT
CLASS="LITERAL"
>GROUP BY</TT
> clause is specified, the
output is combined into groups of rows that match on one or more
values. If the <TT
CLASS="LITERAL"
>HAVING</TT
> clause is present, it
eliminates groups that do not satisfy the given condition. (See
<A
HREF="sql-select.html#SQL-GROUPBY"
><I
><I
>GROUP BY</I
> Clause</I
></A
> and
<A
HREF="sql-select.html#SQL-HAVING"
><I
><I
>HAVING</I
> Clause</I
></A
> below.)
</P
></LI
><LI
><P
> The actual output rows are computed using the
<TT
CLASS="COMMAND"
>SELECT</TT
> output expressions for each selected
row or row group. (See
<A
HREF="sql-select.html#SQL-SELECT-LIST"
><I
><I
>SELECT</I
> List</I
></A
>
below.)
</P
></LI
><LI
><P
><TT
CLASS="LITERAL"
>SELECT DISTINCT</TT
> eliminates duplicate rows from the
result. <TT
CLASS="LITERAL"
>SELECT DISTINCT ON</TT
> eliminates rows that
match on all the specified expressions. <TT
CLASS="LITERAL"
>SELECT ALL</TT
>
(the default) will return all candidate rows, including
duplicates. (See <A
HREF="sql-select.html#SQL-DISTINCT"
><I
><I
>DISTINCT</I
> Clause</I
></A
> below.)
</P
></LI
><LI
><P
> Using the operators <TT
CLASS="LITERAL"
>UNION</TT
>,
<TT
CLASS="LITERAL"
>INTERSECT</TT
>, and <TT
CLASS="LITERAL"
>EXCEPT</TT
>, the
output of more than one <TT
CLASS="COMMAND"
>SELECT</TT
> statement can
be combined to form a single result set. The
<TT
CLASS="LITERAL"
>UNION</TT
> operator returns all rows that are in
one or both of the result sets. The
<TT
CLASS="LITERAL"
>INTERSECT</TT
> operator returns all rows that are
strictly in both result sets. The <TT
CLASS="LITERAL"
>EXCEPT</TT
>
operator returns the rows that are in the first result set but
not in the second. In all three cases, duplicate rows are
eliminated unless <TT
CLASS="LITERAL"
>ALL</TT
> is specified. The noise
word <TT
CLASS="LITERAL"
>DISTINCT</TT
> can be added to explicitly specify
eliminating duplicate rows. Notice that <TT
CLASS="LITERAL"
>DISTINCT</TT
> is
the default behavior here, even though <TT
CLASS="LITERAL"
>ALL</TT
> is
the default for <TT
CLASS="COMMAND"
>SELECT</TT
> itself. (See
<A
HREF="sql-select.html#SQL-UNION"
><I
><I
>UNION</I
> Clause</I
></A
>, <A
HREF="sql-select.html#SQL-INTERSECT"
><I
><I
>INTERSECT</I
> Clause</I
></A
>, and
<A
HREF="sql-select.html#SQL-EXCEPT"
><I
><I
>EXCEPT</I
> Clause</I
></A
> below.)
</P
></LI
><LI
><P
> If the <TT
CLASS="LITERAL"
>ORDER BY</TT
> clause is specified, the
returned rows are sorted in the specified order. If
<TT
CLASS="LITERAL"
>ORDER BY</TT
> is not given, the rows are returned
in whatever order the system finds fastest to produce. (See
<A
HREF="sql-select.html#SQL-ORDERBY"
><I
><I
>ORDER BY</I
> Clause</I
></A
> below.)
</P
></LI
><LI
><P
> If the <TT
CLASS="LITERAL"
>LIMIT</TT
> (or <TT
CLASS="LITERAL"
>FETCH FIRST</TT
>) or <TT
CLASS="LITERAL"
>OFFSET</TT
>
clause is specified, the <TT
CLASS="COMMAND"
>SELECT</TT
> statement
only returns a subset of the result rows. (See <A
HREF="sql-select.html#SQL-LIMIT"
><I
><I
>LIMIT</I
> Clause</I
></A
> below.)
</P
></LI
><LI
><P
> If <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> or <TT
CLASS="LITERAL"
>FOR SHARE</TT
>
is specified, the
<TT
CLASS="COMMAND"
>SELECT</TT
> statement locks the selected rows
against concurrent updates. (See <A
HREF="sql-select.html#SQL-FOR-UPDATE-SHARE"
><I
><I
>FOR UPDATE</I
>/<I
>FOR SHARE</I
> Clause</I
></A
> below.)
</P
></LI
></OL
><P>
</P
><P
> You must have <TT
CLASS="LITERAL"
>SELECT</TT
> privilege on each column used
in a <TT
CLASS="COMMAND"
>SELECT</TT
> command. The use of <TT
CLASS="LITERAL"
>FOR UPDATE</TT
>
or <TT
CLASS="LITERAL"
>FOR SHARE</TT
> requires
<TT
CLASS="LITERAL"
>UPDATE</TT
> privilege as well (for at least one column
of each table so selected).
</P
></DIV
><DIV
CLASS="REFSECT1"
><A
NAME="AEN77728"
></A
><H2
>Parameters</H2
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-WITH"
></A
><H3
><TT
CLASS="LITERAL"
>WITH</TT
> Clause</H3
><P
> The <TT
CLASS="LITERAL"
>WITH</TT
> clause allows you to specify one or more
subqueries that can be referenced by name in the primary query.
The subqueries effectively act as temporary tables or views
for the duration of the primary query.
Each subquery can be a <TT
CLASS="COMMAND"
>SELECT</TT
>, <TT
CLASS="COMMAND"
>VALUES</TT
>,
<TT
CLASS="COMMAND"
>INSERT</TT
>, <TT
CLASS="COMMAND"
>UPDATE</TT
> or
<TT
CLASS="COMMAND"
>DELETE</TT
> statement.
When writing a data-modifying statement (<TT
CLASS="COMMAND"
>INSERT</TT
>,
<TT
CLASS="COMMAND"
>UPDATE</TT
> or <TT
CLASS="COMMAND"
>DELETE</TT
>) in
<TT
CLASS="LITERAL"
>WITH</TT
>, it is usual to include a <TT
CLASS="LITERAL"
>RETURNING</TT
> clause.
It is the output of <TT
CLASS="LITERAL"
>RETURNING</TT
>, <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>not</I
></SPAN
> the underlying
table that the statement modifies, that forms the temporary table that is
read by the primary query. If <TT
CLASS="LITERAL"
>RETURNING</TT
> is omitted, the
statement is still executed, but it produces no output so it cannot be
referenced as a table by the primary query.
</P
><P
> A name (without schema qualification) must be specified for each
<TT
CLASS="LITERAL"
>WITH</TT
> query. Optionally, a list of column names
can be specified; if this is omitted,
the column names are inferred from the subquery.
</P
><P
> If <TT
CLASS="LITERAL"
>RECURSIVE</TT
> is specified, it allows a
<TT
CLASS="COMMAND"
>SELECT</TT
> subquery to reference itself by name. Such a
subquery must have the form
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>non_recursive_term</I
></TT
> UNION [ ALL | DISTINCT ] <TT
CLASS="REPLACEABLE"
><I
>recursive_term</I
></TT
></PRE
><P>
where the recursive self-reference must appear on the right-hand
side of the <TT
CLASS="LITERAL"
>UNION</TT
>. Only one recursive self-reference
is permitted per query. Recursive data-modifying statements are not
supported, but you can use the results of a recursive
<TT
CLASS="COMMAND"
>SELECT</TT
> query in
a data-modifying statement. See <A
HREF="queries-with.html"
>Section 7.8</A
> for
an example.
</P
><P
> Another effect of <TT
CLASS="LITERAL"
>RECURSIVE</TT
> is that
<TT
CLASS="LITERAL"
>WITH</TT
> queries need not be ordered: a query
can reference another one that is later in the list. (However,
circular references, or mutual recursion, are not implemented.)
Without <TT
CLASS="LITERAL"
>RECURSIVE</TT
>, <TT
CLASS="LITERAL"
>WITH</TT
> queries
can only reference sibling <TT
CLASS="LITERAL"
>WITH</TT
> queries
that are earlier in the <TT
CLASS="LITERAL"
>WITH</TT
> list.
</P
><P
> A key property of <TT
CLASS="LITERAL"
>WITH</TT
> queries is that they
are evaluated only once per execution of the primary query,
even if the primary query refers to them more than once.
In particular, data-modifying statements are guaranteed to be
executed once and only once, regardless of whether the primary query
reads all or any of their output.
</P
><P
> The primary query and the <TT
CLASS="LITERAL"
>WITH</TT
> queries are all
(notionally) executed at the same time. This implies that the effects of
a data-modifying statement in <TT
CLASS="LITERAL"
>WITH</TT
> cannot be seen from
other parts of the query, other than by reading its <TT
CLASS="LITERAL"
>RETURNING</TT
>
output. If two such data-modifying statements attempt to modify the same
row, the results are unspecified.
</P
><P
> See <A
HREF="queries-with.html"
>Section 7.8</A
> for additional information.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-FROM"
></A
><H3
><TT
CLASS="LITERAL"
>FROM</TT
> Clause</H3
><P
> The <TT
CLASS="LITERAL"
>FROM</TT
> clause specifies one or more source
tables for the <TT
CLASS="COMMAND"
>SELECT</TT
>. If multiple sources are
specified, the result is the Cartesian product (cross join) of all
the sources. But usually qualification conditions
are added to restrict the returned rows to a small subset of the
Cartesian product.
</P
><P
> The <TT
CLASS="LITERAL"
>FROM</TT
> clause can contain the following
elements:
<P
></P
></P><DIV
CLASS="VARIABLELIST"
><DL
><DT
><TT
CLASS="REPLACEABLE"
><I
>table_name</I
></TT
></DT
><DD
><P
> The name (optionally schema-qualified) of an existing table or view.
If <TT
CLASS="LITERAL"
>ONLY</TT
> is specified before the table name, only that
table is scanned. If <TT
CLASS="LITERAL"
>ONLY</TT
> is not specified, the table
and all its descendant tables (if any) are scanned. Optionally,
<TT
CLASS="LITERAL"
>*</TT
> can be specified after the table name to explicitly
indicate that descendant tables are included.
</P
></DD
><DT
><TT
CLASS="REPLACEABLE"
><I
>alias</I
></TT
></DT
><DD
><P
> A substitute name for the <TT
CLASS="LITERAL"
>FROM</TT
> item containing the
alias. An alias is used for brevity or to eliminate ambiguity
for self-joins (where the same table is scanned multiple
times). When an alias is provided, it completely hides the
actual name of the table or function; for example given
<TT
CLASS="LITERAL"
>FROM foo AS f</TT
>, the remainder of the
<TT
CLASS="COMMAND"
>SELECT</TT
> must refer to this <TT
CLASS="LITERAL"
>FROM</TT
>
item as <TT
CLASS="LITERAL"
>f</TT
> not <TT
CLASS="LITERAL"
>foo</TT
>. If an alias is
written, a column alias list can also be written to provide
substitute names for one or more columns of the table.
</P
></DD
><DT
><TT
CLASS="REPLACEABLE"
><I
>select</I
></TT
></DT
><DD
><P
> A sub-<TT
CLASS="COMMAND"
>SELECT</TT
> can appear in the
<TT
CLASS="LITERAL"
>FROM</TT
> clause. This acts as though its
output were created as a temporary table for the duration of
this single <TT
CLASS="COMMAND"
>SELECT</TT
> command. Note that the
sub-<TT
CLASS="COMMAND"
>SELECT</TT
> must be surrounded by
parentheses, and an alias <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>must</I
></SPAN
> be
provided for it. A
<A
HREF="sql-values.html"
>VALUES</A
> command
can also be used here.
</P
></DD
><DT
><TT
CLASS="REPLACEABLE"
><I
>with_query_name</I
></TT
></DT
><DD
><P
> A <TT
CLASS="LITERAL"
>WITH</TT
> query is referenced by writing its name,
just as though the query's name were a table name. (In fact,
the <TT
CLASS="LITERAL"
>WITH</TT
> query hides any real table of the same name
for the purposes of the primary query. If necessary, you can
refer to a real table of the same name by schema-qualifying
the table's name.)
An alias can be provided in the same way as for a table.
</P
></DD
><DT
><TT
CLASS="REPLACEABLE"
><I
>function_name</I
></TT
></DT
><DD
><P
> Function calls can appear in the <TT
CLASS="LITERAL"
>FROM</TT
>
clause. (This is especially useful for functions that return
result sets, but any function can be used.) This acts as
though its output were created as a temporary table for the
duration of this single <TT
CLASS="COMMAND"
>SELECT</TT
> command. An
alias can also be used. If an alias is written, a column alias
list can also be written to provide substitute names for one
or more attributes of the function's composite return type. If
the function has been defined as returning the <TT
CLASS="TYPE"
>record</TT
>
data type, then an alias or the key word <TT
CLASS="LITERAL"
>AS</TT
> must
be present, followed by a column definition list in the form
<TT
CLASS="LITERAL"
>( <TT
CLASS="REPLACEABLE"
><I
>column_name</I
></TT
> <TT
CLASS="REPLACEABLE"
><I
>data_type</I
></TT
> [<SPAN
CLASS="OPTIONAL"
>, ... </SPAN
>]
)</TT
>. The column definition list must match the actual
number and types of columns returned by the function.
</P
></DD
><DT
><TT
CLASS="REPLACEABLE"
><I
>join_type</I
></TT
></DT
><DD
><P
> One of
<P
></P
></P><UL
><LI
><P
><TT
CLASS="LITERAL"
>[ INNER ] JOIN</TT
></P
></LI
><LI
><P
><TT
CLASS="LITERAL"
>LEFT [ OUTER ] JOIN</TT
></P
></LI
><LI
><P
><TT
CLASS="LITERAL"
>RIGHT [ OUTER ] JOIN</TT
></P
></LI
><LI
><P
><TT
CLASS="LITERAL"
>FULL [ OUTER ] JOIN</TT
></P
></LI
><LI
><P
><TT
CLASS="LITERAL"
>CROSS JOIN</TT
></P
></LI
></UL
><P>
For the <TT
CLASS="LITERAL"
>INNER</TT
> and <TT
CLASS="LITERAL"
>OUTER</TT
> join types, a
join condition must be specified, namely exactly one of
<TT
CLASS="LITERAL"
>NATURAL</TT
>, <TT
CLASS="LITERAL"
>ON <TT
CLASS="REPLACEABLE"
><I
>join_condition</I
></TT
></TT
>, or
<TT
CLASS="LITERAL"
>USING (<TT
CLASS="REPLACEABLE"
><I
>join_column</I
></TT
> [, ...])</TT
>.
See below for the meaning. For <TT
CLASS="LITERAL"
>CROSS JOIN</TT
>,
none of these clauses can appear.
</P
><P
> A <TT
CLASS="LITERAL"
>JOIN</TT
> clause combines two
<TT
CLASS="LITERAL"
>FROM</TT
> items. Use parentheses if necessary to
determine the order of nesting. In the absence of parentheses,
<TT
CLASS="LITERAL"
>JOIN</TT
>s nest left-to-right. In any case
<TT
CLASS="LITERAL"
>JOIN</TT
> binds more tightly than the commas
separating <TT
CLASS="LITERAL"
>FROM</TT
> items.
</P
><P
><TT
CLASS="LITERAL"
>CROSS JOIN</TT
> and <TT
CLASS="LITERAL"
>INNER JOIN</TT
>
produce a simple Cartesian product, the same result as you get from
listing the two items at the top level of <TT
CLASS="LITERAL"
>FROM</TT
>,
but restricted by the join condition (if any).
<TT
CLASS="LITERAL"
>CROSS JOIN</TT
> is equivalent to <TT
CLASS="LITERAL"
>INNER JOIN ON
(TRUE)</TT
>, that is, no rows are removed by qualification.
These join types are just a notational convenience, since they
do nothing you couldn't do with plain <TT
CLASS="LITERAL"
>FROM</TT
> and
<TT
CLASS="LITERAL"
>WHERE</TT
>.
</P
><P
><TT
CLASS="LITERAL"
>LEFT OUTER JOIN</TT
> returns all rows in the qualified
Cartesian product (i.e., all combined rows that pass its join
condition), plus one copy of each row in the left-hand table
for which there was no right-hand row that passed the join
condition. This left-hand row is extended to the full width
of the joined table by inserting null values for the
right-hand columns. Note that only the <TT
CLASS="LITERAL"
>JOIN</TT
>
clause's own condition is considered while deciding which rows
have matches. Outer conditions are applied afterwards.
</P
><P
> Conversely, <TT
CLASS="LITERAL"
>RIGHT OUTER JOIN</TT
> returns all the
joined rows, plus one row for each unmatched right-hand row
(extended with nulls on the left). This is just a notational
convenience, since you could convert it to a <TT
CLASS="LITERAL"
>LEFT
OUTER JOIN</TT
> by switching the left and right inputs.
</P
><P
><TT
CLASS="LITERAL"
>FULL OUTER JOIN</TT
> returns all the joined rows, plus
one row for each unmatched left-hand row (extended with nulls
on the right), plus one row for each unmatched right-hand row
(extended with nulls on the left).
</P
></DD
><DT
><TT
CLASS="LITERAL"
>ON <TT
CLASS="REPLACEABLE"
><I
>join_condition</I
></TT
></TT
></DT
><DD
><P
><TT
CLASS="REPLACEABLE"
><I
>join_condition</I
></TT
> is
an expression resulting in a value of type
<TT
CLASS="TYPE"
>boolean</TT
> (similar to a <TT
CLASS="LITERAL"
>WHERE</TT
>
clause) that specifies which rows in a join are considered to
match.
</P
></DD
><DT
><TT
CLASS="LITERAL"
>USING ( <TT
CLASS="REPLACEABLE"
><I
>join_column</I
></TT
> [, ...] )</TT
></DT
><DD
><P
> A clause of the form <TT
CLASS="LITERAL"
>USING ( a, b, ... )</TT
> is
shorthand for <TT
CLASS="LITERAL"
>ON left_table.a = right_table.a AND
left_table.b = right_table.b ...</TT
>. Also,
<TT
CLASS="LITERAL"
>USING</TT
> implies that only one of each pair of
equivalent columns will be included in the join output, not
both.
</P
></DD
><DT
><TT
CLASS="LITERAL"
>NATURAL</TT
></DT
><DD
><P
> <TT
CLASS="LITERAL"
>NATURAL</TT
> is shorthand for a
<TT
CLASS="LITERAL"
>USING</TT
> list that mentions all columns in the two
tables that have matching names. If there are no common
column names, <TT
CLASS="LITERAL"
>NATURAL</TT
> is equivalent
to <TT
CLASS="LITERAL"
>ON TRUE</TT
>.
</P
></DD
></DL
></DIV
><P>
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-WHERE"
></A
><H3
><TT
CLASS="LITERAL"
>WHERE</TT
> Clause</H3
><P
> The optional <TT
CLASS="LITERAL"
>WHERE</TT
> clause has the general form
</P><PRE
CLASS="SYNOPSIS"
>WHERE <TT
CLASS="REPLACEABLE"
><I
>condition</I
></TT
></PRE
><P>
where <TT
CLASS="REPLACEABLE"
><I
>condition</I
></TT
> is
any expression that evaluates to a result of type
<TT
CLASS="TYPE"
>boolean</TT
>. Any row that does not satisfy this
condition will be eliminated from the output. A row satisfies the
condition if it returns true when the actual row values are
substituted for any variable references.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-GROUPBY"
></A
><H3
><TT
CLASS="LITERAL"
>GROUP BY</TT
> Clause</H3
><P
> The optional <TT
CLASS="LITERAL"
>GROUP BY</TT
> clause has the general form
</P><PRE
CLASS="SYNOPSIS"
>GROUP BY <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [, ...]</PRE
><P>
</P
><P
> <TT
CLASS="LITERAL"
>GROUP BY</TT
> will condense into a single row all
selected rows that share the same values for the grouped
expressions. <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> can be an input column
name, or the name or ordinal number of an output column
(<TT
CLASS="COMMAND"
>SELECT</TT
> list item), or an arbitrary
expression formed from input-column values. In case of ambiguity,
a <TT
CLASS="LITERAL"
>GROUP BY</TT
> name will be interpreted as an
input-column name rather than an output column name.
</P
><P
> Aggregate functions, if any are used, are computed across all rows
making up each group, producing a separate value for each group
(whereas without <TT
CLASS="LITERAL"
>GROUP BY</TT
>, an aggregate
produces a single value computed across all the selected rows).
When <TT
CLASS="LITERAL"
>GROUP BY</TT
> is present, it is not valid for
the <TT
CLASS="COMMAND"
>SELECT</TT
> list expressions to refer to
ungrouped columns except within aggregate functions or if the
ungrouped column is functionally dependent on the grouped columns,
since there would otherwise be more than one possible value to
return for an ungrouped column. A functional dependency exists if
the grouped columns (or a subset thereof) are the primary key of
the table containing the ungrouped column.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-HAVING"
></A
><H3
><TT
CLASS="LITERAL"
>HAVING</TT
> Clause</H3
><P
> The optional <TT
CLASS="LITERAL"
>HAVING</TT
> clause has the general form
</P><PRE
CLASS="SYNOPSIS"
>HAVING <TT
CLASS="REPLACEABLE"
><I
>condition</I
></TT
></PRE
><P>
where <TT
CLASS="REPLACEABLE"
><I
>condition</I
></TT
> is
the same as specified for the <TT
CLASS="LITERAL"
>WHERE</TT
> clause.
</P
><P
> <TT
CLASS="LITERAL"
>HAVING</TT
> eliminates group rows that do not
satisfy the condition. <TT
CLASS="LITERAL"
>HAVING</TT
> is different
from <TT
CLASS="LITERAL"
>WHERE</TT
>: <TT
CLASS="LITERAL"
>WHERE</TT
> filters
individual rows before the application of <TT
CLASS="LITERAL"
>GROUP
BY</TT
>, while <TT
CLASS="LITERAL"
>HAVING</TT
> filters group rows
created by <TT
CLASS="LITERAL"
>GROUP BY</TT
>. Each column referenced in
<TT
CLASS="REPLACEABLE"
><I
>condition</I
></TT
> must
unambiguously reference a grouping column, unless the reference
appears within an aggregate function.
</P
><P
> The presence of <TT
CLASS="LITERAL"
>HAVING</TT
> turns a query into a grouped
query even if there is no <TT
CLASS="LITERAL"
>GROUP BY</TT
> clause. This is the
same as what happens when the query contains aggregate functions but
no <TT
CLASS="LITERAL"
>GROUP BY</TT
> clause. All the selected rows are considered to
form a single group, and the <TT
CLASS="COMMAND"
>SELECT</TT
> list and
<TT
CLASS="LITERAL"
>HAVING</TT
> clause can only reference table columns from
within aggregate functions. Such a query will emit a single row if the
<TT
CLASS="LITERAL"
>HAVING</TT
> condition is true, zero rows if it is not true.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-WINDOW"
></A
><H3
><TT
CLASS="LITERAL"
>WINDOW</TT
> Clause</H3
><P
> The optional <TT
CLASS="LITERAL"
>WINDOW</TT
> clause has the general form
</P><PRE
CLASS="SYNOPSIS"
>WINDOW <TT
CLASS="REPLACEABLE"
><I
>window_name</I
></TT
> AS ( <TT
CLASS="REPLACEABLE"
><I
>window_definition</I
></TT
> ) [, ...]</PRE
><P>
where <TT
CLASS="REPLACEABLE"
><I
>window_name</I
></TT
> is
a name that can be referenced from <TT
CLASS="LITERAL"
>OVER</TT
> clauses or
subsequent window definitions, and
<TT
CLASS="REPLACEABLE"
><I
>window_definition</I
></TT
> is
</P><PRE
CLASS="SYNOPSIS"
>[ <TT
CLASS="REPLACEABLE"
><I
>existing_window_name</I
></TT
> ]
[ PARTITION BY <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [, ...] ]
[ ORDER BY <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [ ASC | DESC | USING <TT
CLASS="REPLACEABLE"
><I
>operator</I
></TT
> ] [ NULLS { FIRST | LAST } ] [, ...] ]
[ <TT
CLASS="REPLACEABLE"
><I
>frame_clause</I
></TT
> ]</PRE
><P>
</P
><P
> If an <TT
CLASS="REPLACEABLE"
><I
>existing_window_name</I
></TT
>
is specified it must refer to an earlier entry in the <TT
CLASS="LITERAL"
>WINDOW</TT
>
list; the new window copies its partitioning clause from that entry,
as well as its ordering clause if any. In this case the new window cannot
specify its own <TT
CLASS="LITERAL"
>PARTITION BY</TT
> clause, and it can specify
<TT
CLASS="LITERAL"
>ORDER BY</TT
> only if the copied window does not have one.
The new window always uses its own frame clause; the copied window
must not specify a frame clause.
</P
><P
> The elements of the <TT
CLASS="LITERAL"
>PARTITION BY</TT
> list are interpreted in
much the same fashion as elements of a
<A
HREF="sql-select.html#SQL-GROUPBY"
><I
><I
>GROUP BY</I
> Clause</I
></A
>, except that
they are always simple expressions and never the name or number of an
output column.
Another difference is that these expressions can contain aggregate
function calls, which are not allowed in a regular <TT
CLASS="LITERAL"
>GROUP BY</TT
>
clause. They are allowed here because windowing occurs after grouping
and aggregation.
</P
><P
> Similarly, the elements of the <TT
CLASS="LITERAL"
>ORDER BY</TT
> list are interpreted
in much the same fashion as elements of an
<A
HREF="sql-select.html#SQL-ORDERBY"
><I
><I
>ORDER BY</I
> Clause</I
></A
>, except that
the expressions are always taken as simple expressions and never the name
or number of an output column.
</P
><P
> The optional <TT
CLASS="REPLACEABLE"
><I
>frame_clause</I
></TT
> defines
the <I
CLASS="FIRSTTERM"
>window frame</I
> for window functions that depend on the
frame (not all do). The window frame is a set of related rows for
each row of the query (called the <I
CLASS="FIRSTTERM"
>current row</I
>).
The <TT
CLASS="REPLACEABLE"
><I
>frame_clause</I
></TT
> can be one of
</P><PRE
CLASS="SYNOPSIS"
>{ RANGE | ROWS } <TT
CLASS="REPLACEABLE"
><I
>frame_start</I
></TT
>
{ RANGE | ROWS } BETWEEN <TT
CLASS="REPLACEABLE"
><I
>frame_start</I
></TT
> AND <TT
CLASS="REPLACEABLE"
><I
>frame_end</I
></TT
></PRE
><P>
where <TT
CLASS="REPLACEABLE"
><I
>frame_start</I
></TT
> and <TT
CLASS="REPLACEABLE"
><I
>frame_end</I
></TT
> can be
one of
</P><PRE
CLASS="SYNOPSIS"
>UNBOUNDED PRECEDING
<TT
CLASS="REPLACEABLE"
><I
>value</I
></TT
> PRECEDING
CURRENT ROW
<TT
CLASS="REPLACEABLE"
><I
>value</I
></TT
> FOLLOWING
UNBOUNDED FOLLOWING</PRE
><P>
If <TT
CLASS="REPLACEABLE"
><I
>frame_end</I
></TT
> is omitted it defaults to <TT
CLASS="LITERAL"
>CURRENT
ROW</TT
>. Restrictions are that
<TT
CLASS="REPLACEABLE"
><I
>frame_start</I
></TT
> cannot be <TT
CLASS="LITERAL"
>UNBOUNDED FOLLOWING</TT
>,
<TT
CLASS="REPLACEABLE"
><I
>frame_end</I
></TT
> cannot be <TT
CLASS="LITERAL"
>UNBOUNDED PRECEDING</TT
>,
and the <TT
CLASS="REPLACEABLE"
><I
>frame_end</I
></TT
> choice cannot appear earlier in the
above list than the <TT
CLASS="REPLACEABLE"
><I
>frame_start</I
></TT
> choice — for example
<TT
CLASS="LITERAL"
>RANGE BETWEEN CURRENT ROW AND <TT
CLASS="REPLACEABLE"
><I
>value</I
></TT
>
PRECEDING</TT
> is not allowed.
</P
><P
> The default framing option is <TT
CLASS="LITERAL"
>RANGE UNBOUNDED PRECEDING</TT
>,
which is the same as <TT
CLASS="LITERAL"
>RANGE BETWEEN UNBOUNDED PRECEDING AND
CURRENT ROW</TT
>; it sets the frame to be all rows from the partition start
up through the current row's last peer in the <TT
CLASS="LITERAL"
>ORDER BY</TT
>
ordering (which means all rows if there is no <TT
CLASS="LITERAL"
>ORDER BY</TT
>).
In general, <TT
CLASS="LITERAL"
>UNBOUNDED PRECEDING</TT
> means that the frame
starts with the first row of the partition, and similarly
<TT
CLASS="LITERAL"
>UNBOUNDED FOLLOWING</TT
> means that the frame ends with the last
row of the partition (regardless of <TT
CLASS="LITERAL"
>RANGE</TT
> or <TT
CLASS="LITERAL"
>ROWS</TT
>
mode). In <TT
CLASS="LITERAL"
>ROWS</TT
> mode, <TT
CLASS="LITERAL"
>CURRENT ROW</TT
>
means that the frame starts or ends with the current row; but in
<TT
CLASS="LITERAL"
>RANGE</TT
> mode it means that the frame starts or ends with
the current row's first or last peer in the <TT
CLASS="LITERAL"
>ORDER BY</TT
> ordering.
The <TT
CLASS="REPLACEABLE"
><I
>value</I
></TT
> <TT
CLASS="LITERAL"
>PRECEDING</TT
> and
<TT
CLASS="REPLACEABLE"
><I
>value</I
></TT
> <TT
CLASS="LITERAL"
>FOLLOWING</TT
> cases are currently only
allowed in <TT
CLASS="LITERAL"
>ROWS</TT
> mode. They indicate that the frame starts
or ends with the row that many rows before or after the current row.
<TT
CLASS="REPLACEABLE"
><I
>value</I
></TT
> must be an integer expression not
containing any variables, aggregate functions, or window functions.
The value must not be null or negative; but it can be zero, which
selects the current row itself.
</P
><P
> Beware that the <TT
CLASS="LITERAL"
>ROWS</TT
> options can produce unpredictable
results if the <TT
CLASS="LITERAL"
>ORDER BY</TT
> ordering does not order the rows
uniquely. The <TT
CLASS="LITERAL"
>RANGE</TT
> options are designed to ensure that
rows that are peers in the <TT
CLASS="LITERAL"
>ORDER BY</TT
> ordering are treated
alike; any two peer rows will be both in or both not in the frame.
</P
><P
> The purpose of a <TT
CLASS="LITERAL"
>WINDOW</TT
> clause is to specify the
behavior of <I
CLASS="FIRSTTERM"
>window functions</I
> appearing in the query's
<A
HREF="sql-select.html#SQL-SELECT-LIST"
><I
><I
>SELECT</I
> List</I
></A
> or
<A
HREF="sql-select.html#SQL-ORDERBY"
><I
><I
>ORDER BY</I
> Clause</I
></A
>. These functions
can reference the <TT
CLASS="LITERAL"
>WINDOW</TT
> clause entries by name
in their <TT
CLASS="LITERAL"
>OVER</TT
> clauses. A <TT
CLASS="LITERAL"
>WINDOW</TT
> clause
entry does not have to be referenced anywhere, however; if it is not
used in the query it is simply ignored. It is possible to use window
functions without any <TT
CLASS="LITERAL"
>WINDOW</TT
> clause at all, since
a window function call can specify its window definition directly in
its <TT
CLASS="LITERAL"
>OVER</TT
> clause. However, the <TT
CLASS="LITERAL"
>WINDOW</TT
>
clause saves typing when the same window definition is needed for more
than one window function.
</P
><P
> Window functions are described in detail in
<A
HREF="tutorial-window.html"
>Section 3.5</A
>,
<A
HREF="sql-expressions.html#SYNTAX-WINDOW-FUNCTIONS"
>Section 4.2.8</A
>, and
<A
HREF="queries-table-expressions.html#QUERIES-WINDOW"
>Section 7.2.4</A
>.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-SELECT-LIST"
></A
><H3
><TT
CLASS="COMMAND"
>SELECT</TT
> List</H3
><P
> The <TT
CLASS="COMMAND"
>SELECT</TT
> list (between the key words
<TT
CLASS="LITERAL"
>SELECT</TT
> and <TT
CLASS="LITERAL"
>FROM</TT
>) specifies expressions
that form the output rows of the <TT
CLASS="COMMAND"
>SELECT</TT
>
statement. The expressions can (and usually do) refer to columns
computed in the <TT
CLASS="LITERAL"
>FROM</TT
> clause.
</P
><P
> Just as in a table, every output column of a <TT
CLASS="COMMAND"
>SELECT</TT
>
has a name. In a simple <TT
CLASS="COMMAND"
>SELECT</TT
> this name is just
used to label the column for display, but when the <TT
CLASS="COMMAND"
>SELECT</TT
>
is a sub-query of a larger query, the name is seen by the larger query
as the column name of the virtual table produced by the sub-query.
To specify the name to use for an output column, write
<TT
CLASS="LITERAL"
>AS</TT
> <TT
CLASS="REPLACEABLE"
><I
>output_name</I
></TT
>
after the column's expression. (You can omit <TT
CLASS="LITERAL"
>AS</TT
>,
but only if the desired output name does not match any
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> keyword (see <A
HREF="sql-keywords-appendix.html"
>Appendix C</A
>). For protection against possible
future keyword additions, it is recommended that you always either
write <TT
CLASS="LITERAL"
>AS</TT
> or double-quote the output name.)
If you do not specify a column name, a name is chosen automatically
by <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>. If the column's expression
is a simple column reference then the chosen name is the same as that
column's name. In more complex cases a function or type name may be
used, or the system may fall back on a generated name such as
<TT
CLASS="LITERAL"
>?column?</TT
>.
</P
><P
> An output column's name can be used to refer to the column's value in
<TT
CLASS="LITERAL"
>ORDER BY</TT
> and <TT
CLASS="LITERAL"
>GROUP BY</TT
> clauses, but not in the
<TT
CLASS="LITERAL"
>WHERE</TT
> or <TT
CLASS="LITERAL"
>HAVING</TT
> clauses; there you must write
out the expression instead.
</P
><P
> Instead of an expression, <TT
CLASS="LITERAL"
>*</TT
> can be written in
the output list as a shorthand for all the columns of the selected
rows. Also, you can write <TT
CLASS="LITERAL"
><TT
CLASS="REPLACEABLE"
><I
>table_name</I
></TT
>.*</TT
> as a
shorthand for the columns coming from just that table. In these
cases it is not possible to specify new names with <TT
CLASS="LITERAL"
>AS</TT
>;
the output column names will be the same as the table columns' names.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-DISTINCT"
></A
><H3
><TT
CLASS="LITERAL"
>DISTINCT</TT
> Clause</H3
><P
> If <TT
CLASS="LITERAL"
>SELECT DISTINCT</TT
> is specified, all duplicate rows are
removed from the result set (one row is kept from each group of
duplicates). <TT
CLASS="LITERAL"
>SELECT ALL</TT
> specifies the opposite: all rows are
kept; that is the default.
</P
><P
> <TT
CLASS="LITERAL"
>SELECT DISTINCT ON ( <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [, ...] )</TT
>
keeps only the first row of each set of rows where the given
expressions evaluate to equal. The <TT
CLASS="LITERAL"
>DISTINCT ON</TT
>
expressions are interpreted using the same rules as for
<TT
CLASS="LITERAL"
>ORDER BY</TT
> (see above). Note that the <SPAN
CLASS="QUOTE"
>"first
row"</SPAN
> of each set is unpredictable unless <TT
CLASS="LITERAL"
>ORDER
BY</TT
> is used to ensure that the desired row appears first. For
example:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT DISTINCT ON (location) location, time, report
FROM weather_reports
ORDER BY location, time DESC;</PRE
><P>
retrieves the most recent weather report for each location. But
if we had not used <TT
CLASS="LITERAL"
>ORDER BY</TT
> to force descending order
of time values for each location, we'd have gotten a report from
an unpredictable time for each location.
</P
><P
> The <TT
CLASS="LITERAL"
>DISTINCT ON</TT
> expression(s) must match the leftmost
<TT
CLASS="LITERAL"
>ORDER BY</TT
> expression(s). The <TT
CLASS="LITERAL"
>ORDER BY</TT
> clause
will normally contain additional expression(s) that determine the
desired precedence of rows within each <TT
CLASS="LITERAL"
>DISTINCT ON</TT
> group.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-UNION"
></A
><H3
><TT
CLASS="LITERAL"
>UNION</TT
> Clause</H3
><P
> The <TT
CLASS="LITERAL"
>UNION</TT
> clause has this general form:
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>select_statement</I
></TT
> UNION [ ALL | DISTINCT ] <TT
CLASS="REPLACEABLE"
><I
>select_statement</I
></TT
></PRE
><P><TT
CLASS="REPLACEABLE"
><I
>select_statement</I
></TT
> is
any <TT
CLASS="COMMAND"
>SELECT</TT
> statement without an <TT
CLASS="LITERAL"
>ORDER
BY</TT
>, <TT
CLASS="LITERAL"
>LIMIT</TT
>, <TT
CLASS="LITERAL"
>FOR UPDATE</TT
>, or
<TT
CLASS="LITERAL"
>FOR SHARE</TT
> clause.
(<TT
CLASS="LITERAL"
>ORDER BY</TT
> and <TT
CLASS="LITERAL"
>LIMIT</TT
> can be attached to a
subexpression if it is enclosed in parentheses. Without
parentheses, these clauses will be taken to apply to the result of
the <TT
CLASS="LITERAL"
>UNION</TT
>, not to its right-hand input
expression.)
</P
><P
> The <TT
CLASS="LITERAL"
>UNION</TT
> operator computes the set union of
the rows returned by the involved <TT
CLASS="COMMAND"
>SELECT</TT
>
statements. A row is in the set union of two result sets if it
appears in at least one of the result sets. The two
<TT
CLASS="COMMAND"
>SELECT</TT
> statements that represent the direct
operands of the <TT
CLASS="LITERAL"
>UNION</TT
> must produce the same
number of columns, and corresponding columns must be of compatible
data types.
</P
><P
> The result of <TT
CLASS="LITERAL"
>UNION</TT
> does not contain any duplicate
rows unless the <TT
CLASS="LITERAL"
>ALL</TT
> option is specified.
<TT
CLASS="LITERAL"
>ALL</TT
> prevents elimination of duplicates. (Therefore,
<TT
CLASS="LITERAL"
>UNION ALL</TT
> is usually significantly quicker than
<TT
CLASS="LITERAL"
>UNION</TT
>; use <TT
CLASS="LITERAL"
>ALL</TT
> when you can.)
<TT
CLASS="LITERAL"
>DISTINCT</TT
> can be written to explicitly specify the
default behavior of eliminating duplicate rows.
</P
><P
> Multiple <TT
CLASS="LITERAL"
>UNION</TT
> operators in the same
<TT
CLASS="COMMAND"
>SELECT</TT
> statement are evaluated left to right,
unless otherwise indicated by parentheses.
</P
><P
> Currently, <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> and <TT
CLASS="LITERAL"
>FOR SHARE</TT
> cannot be
specified either for a <TT
CLASS="LITERAL"
>UNION</TT
> result or for any input of a
<TT
CLASS="LITERAL"
>UNION</TT
>.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-INTERSECT"
></A
><H3
><TT
CLASS="LITERAL"
>INTERSECT</TT
> Clause</H3
><P
> The <TT
CLASS="LITERAL"
>INTERSECT</TT
> clause has this general form:
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>select_statement</I
></TT
> INTERSECT [ ALL | DISTINCT ] <TT
CLASS="REPLACEABLE"
><I
>select_statement</I
></TT
></PRE
><P><TT
CLASS="REPLACEABLE"
><I
>select_statement</I
></TT
> is
any <TT
CLASS="COMMAND"
>SELECT</TT
> statement without an <TT
CLASS="LITERAL"
>ORDER
BY</TT
>, <TT
CLASS="LITERAL"
>LIMIT</TT
>, <TT
CLASS="LITERAL"
>FOR UPDATE</TT
>, or
<TT
CLASS="LITERAL"
>FOR SHARE</TT
> clause.
</P
><P
> The <TT
CLASS="LITERAL"
>INTERSECT</TT
> operator computes the set
intersection of the rows returned by the involved
<TT
CLASS="COMMAND"
>SELECT</TT
> statements. A row is in the
intersection of two result sets if it appears in both result sets.
</P
><P
> The result of <TT
CLASS="LITERAL"
>INTERSECT</TT
> does not contain any
duplicate rows unless the <TT
CLASS="LITERAL"
>ALL</TT
> option is specified.
With <TT
CLASS="LITERAL"
>ALL</TT
>, a row that has <TT
CLASS="REPLACEABLE"
><I
>m</I
></TT
> duplicates in the
left table and <TT
CLASS="REPLACEABLE"
><I
>n</I
></TT
> duplicates in the right table will appear
min(<TT
CLASS="REPLACEABLE"
><I
>m</I
></TT
>,<TT
CLASS="REPLACEABLE"
><I
>n</I
></TT
>) times in the result set.
<TT
CLASS="LITERAL"
>DISTINCT</TT
> can be written to explicitly specify the
default behavior of eliminating duplicate rows.
</P
><P
> Multiple <TT
CLASS="LITERAL"
>INTERSECT</TT
> operators in the same
<TT
CLASS="COMMAND"
>SELECT</TT
> statement are evaluated left to right,
unless parentheses dictate otherwise.
<TT
CLASS="LITERAL"
>INTERSECT</TT
> binds more tightly than
<TT
CLASS="LITERAL"
>UNION</TT
>. That is, <TT
CLASS="LITERAL"
>A UNION B INTERSECT
C</TT
> will be read as <TT
CLASS="LITERAL"
>A UNION (B INTERSECT
C)</TT
>.
</P
><P
> Currently, <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> and <TT
CLASS="LITERAL"
>FOR SHARE</TT
> cannot be
specified either for an <TT
CLASS="LITERAL"
>INTERSECT</TT
> result or for any input of
an <TT
CLASS="LITERAL"
>INTERSECT</TT
>.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-EXCEPT"
></A
><H3
><TT
CLASS="LITERAL"
>EXCEPT</TT
> Clause</H3
><P
> The <TT
CLASS="LITERAL"
>EXCEPT</TT
> clause has this general form:
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>select_statement</I
></TT
> EXCEPT [ ALL | DISTINCT ] <TT
CLASS="REPLACEABLE"
><I
>select_statement</I
></TT
></PRE
><P><TT
CLASS="REPLACEABLE"
><I
>select_statement</I
></TT
> is
any <TT
CLASS="COMMAND"
>SELECT</TT
> statement without an <TT
CLASS="LITERAL"
>ORDER
BY</TT
>, <TT
CLASS="LITERAL"
>LIMIT</TT
>, <TT
CLASS="LITERAL"
>FOR UPDATE</TT
>, or
<TT
CLASS="LITERAL"
>FOR SHARE</TT
> clause.
</P
><P
> The <TT
CLASS="LITERAL"
>EXCEPT</TT
> operator computes the set of rows
that are in the result of the left <TT
CLASS="COMMAND"
>SELECT</TT
>
statement but not in the result of the right one.
</P
><P
> The result of <TT
CLASS="LITERAL"
>EXCEPT</TT
> does not contain any
duplicate rows unless the <TT
CLASS="LITERAL"
>ALL</TT
> option is specified.
With <TT
CLASS="LITERAL"
>ALL</TT
>, a row that has <TT
CLASS="REPLACEABLE"
><I
>m</I
></TT
> duplicates in the
left table and <TT
CLASS="REPLACEABLE"
><I
>n</I
></TT
> duplicates in the right table will appear
max(<TT
CLASS="REPLACEABLE"
><I
>m</I
></TT
>-<TT
CLASS="REPLACEABLE"
><I
>n</I
></TT
>,0) times in the result set.
<TT
CLASS="LITERAL"
>DISTINCT</TT
> can be written to explicitly specify the
default behavior of eliminating duplicate rows.
</P
><P
> Multiple <TT
CLASS="LITERAL"
>EXCEPT</TT
> operators in the same
<TT
CLASS="COMMAND"
>SELECT</TT
> statement are evaluated left to right,
unless parentheses dictate otherwise. <TT
CLASS="LITERAL"
>EXCEPT</TT
> binds at
the same level as <TT
CLASS="LITERAL"
>UNION</TT
>.
</P
><P
> Currently, <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> and <TT
CLASS="LITERAL"
>FOR SHARE</TT
> cannot be
specified either for an <TT
CLASS="LITERAL"
>EXCEPT</TT
> result or for any input of
an <TT
CLASS="LITERAL"
>EXCEPT</TT
>.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-ORDERBY"
></A
><H3
><TT
CLASS="LITERAL"
>ORDER BY</TT
> Clause</H3
><P
> The optional <TT
CLASS="LITERAL"
>ORDER BY</TT
> clause has this general form:
</P><PRE
CLASS="SYNOPSIS"
>ORDER BY <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [ ASC | DESC | USING <TT
CLASS="REPLACEABLE"
><I
>operator</I
></TT
> ] [ NULLS { FIRST | LAST } ] [, ...]</PRE
><P>
The <TT
CLASS="LITERAL"
>ORDER BY</TT
> clause causes the result rows to
be sorted according to the specified expression(s). If two rows are
equal according to the leftmost expression, they are compared
according to the next expression and so on. If they are equal
according to all specified expressions, they are returned in
an implementation-dependent order.
</P
><P
> Each <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> can be the
name or ordinal number of an output column
(<TT
CLASS="COMMAND"
>SELECT</TT
> list item), or it can be an arbitrary
expression formed from input-column values.
</P
><P
> The ordinal number refers to the ordinal (left-to-right) position
of the output column. This feature makes it possible to define an
ordering on the basis of a column that does not have a unique
name. This is never absolutely necessary because it is always
possible to assign a name to an output column using the
<TT
CLASS="LITERAL"
>AS</TT
> clause.
</P
><P
> It is also possible to use arbitrary expressions in the
<TT
CLASS="LITERAL"
>ORDER BY</TT
> clause, including columns that do not
appear in the <TT
CLASS="COMMAND"
>SELECT</TT
> output list. Thus the
following statement is valid:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT name FROM distributors ORDER BY code;</PRE
><P>
A limitation of this feature is that an <TT
CLASS="LITERAL"
>ORDER BY</TT
>
clause applying to the result of a <TT
CLASS="LITERAL"
>UNION</TT
>,
<TT
CLASS="LITERAL"
>INTERSECT</TT
>, or <TT
CLASS="LITERAL"
>EXCEPT</TT
> clause can only
specify an output column name or number, not an expression.
</P
><P
> If an <TT
CLASS="LITERAL"
>ORDER BY</TT
> expression is a simple name that
matches both an output column name and an input column name,
<TT
CLASS="LITERAL"
>ORDER BY</TT
> will interpret it as the output column name.
This is the opposite of the choice that <TT
CLASS="LITERAL"
>GROUP BY</TT
> will
make in the same situation. This inconsistency is made to be
compatible with the SQL standard.
</P
><P
> Optionally one can add the key word <TT
CLASS="LITERAL"
>ASC</TT
> (ascending) or
<TT
CLASS="LITERAL"
>DESC</TT
> (descending) after any expression in the
<TT
CLASS="LITERAL"
>ORDER BY</TT
> clause. If not specified, <TT
CLASS="LITERAL"
>ASC</TT
> is
assumed by default. Alternatively, a specific ordering operator
name can be specified in the <TT
CLASS="LITERAL"
>USING</TT
> clause.
An ordering operator must be a less-than or greater-than
member of some B-tree operator family.
<TT
CLASS="LITERAL"
>ASC</TT
> is usually equivalent to <TT
CLASS="LITERAL"
>USING <</TT
> and
<TT
CLASS="LITERAL"
>DESC</TT
> is usually equivalent to <TT
CLASS="LITERAL"
>USING ></TT
>.
(But the creator of a user-defined data type can define exactly what the
default sort ordering is, and it might correspond to operators with other
names.)
</P
><P
> If <TT
CLASS="LITERAL"
>NULLS LAST</TT
> is specified, null values sort after all
non-null values; if <TT
CLASS="LITERAL"
>NULLS FIRST</TT
> is specified, null values
sort before all non-null values. If neither is specified, the default
behavior is <TT
CLASS="LITERAL"
>NULLS LAST</TT
> when <TT
CLASS="LITERAL"
>ASC</TT
> is specified
or implied, and <TT
CLASS="LITERAL"
>NULLS FIRST</TT
> when <TT
CLASS="LITERAL"
>DESC</TT
> is specified
(thus, the default is to act as though nulls are larger than non-nulls).
When <TT
CLASS="LITERAL"
>USING</TT
> is specified, the default nulls ordering depends
on whether the operator is a less-than or greater-than operator.
</P
><P
> Note that ordering options apply only to the expression they follow;
for example <TT
CLASS="LITERAL"
>ORDER BY x, y DESC</TT
> does not mean
the same thing as <TT
CLASS="LITERAL"
>ORDER BY x DESC, y DESC</TT
>.
</P
><P
> Character-string data is sorted according to the collation that applies
to the column being sorted. That can be overridden at need by including
a <TT
CLASS="LITERAL"
>COLLATE</TT
> clause in the
<TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
>, for example
<TT
CLASS="LITERAL"
>ORDER BY mycolumn COLLATE "en_US"</TT
>.
For more information see <A
HREF="sql-expressions.html#SQL-SYNTAX-COLLATE-EXPRS"
>Section 4.2.10</A
> and
<A
HREF="collation.html"
>Section 22.2</A
>.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-LIMIT"
></A
><H3
><TT
CLASS="LITERAL"
>LIMIT</TT
> Clause</H3
><P
> The <TT
CLASS="LITERAL"
>LIMIT</TT
> clause consists of two independent
sub-clauses:
</P><PRE
CLASS="SYNOPSIS"
>LIMIT { <TT
CLASS="REPLACEABLE"
><I
>count</I
></TT
> | ALL }
OFFSET <TT
CLASS="REPLACEABLE"
><I
>start</I
></TT
></PRE
><P>
<TT
CLASS="REPLACEABLE"
><I
>count</I
></TT
> specifies the
maximum number of rows to return, while <TT
CLASS="REPLACEABLE"
><I
>start</I
></TT
> specifies the number of rows
to skip before starting to return rows. When both are specified,
<TT
CLASS="REPLACEABLE"
><I
>start</I
></TT
> rows are skipped
before starting to count the <TT
CLASS="REPLACEABLE"
><I
>count</I
></TT
> rows to be returned.
</P
><P
> If the <TT
CLASS="REPLACEABLE"
><I
>count</I
></TT
> expression
evaluates to NULL, it is treated as <TT
CLASS="LITERAL"
>LIMIT ALL</TT
>, i.e., no
limit. If <TT
CLASS="REPLACEABLE"
><I
>start</I
></TT
> evaluates
to NULL, it is treated the same as <TT
CLASS="LITERAL"
>OFFSET 0</TT
>.
</P
><P
> SQL:2008 introduced a different syntax to achieve the same result,
which <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> also supports. It is:
</P><PRE
CLASS="SYNOPSIS"
>OFFSET <TT
CLASS="REPLACEABLE"
><I
>start</I
></TT
> { ROW | ROWS }
FETCH { FIRST | NEXT } [ <TT
CLASS="REPLACEABLE"
><I
>count</I
></TT
> ] { ROW | ROWS } ONLY</PRE
><P>
In this syntax, to write anything except a simple integer constant for
<TT
CLASS="REPLACEABLE"
><I
>start</I
></TT
> or <TT
CLASS="REPLACEABLE"
><I
>count</I
></TT
>, you must write parentheses
around it.
If <TT
CLASS="REPLACEABLE"
><I
>count</I
></TT
> is
omitted in a <TT
CLASS="LITERAL"
>FETCH</TT
> clause, it defaults to 1.
<TT
CLASS="LITERAL"
>ROW</TT
>
and <TT
CLASS="LITERAL"
>ROWS</TT
> as well as <TT
CLASS="LITERAL"
>FIRST</TT
>
and <TT
CLASS="LITERAL"
>NEXT</TT
> are noise words that don't influence
the effects of these clauses.
According to the standard, the <TT
CLASS="LITERAL"
>OFFSET</TT
> clause must come
before the <TT
CLASS="LITERAL"
>FETCH</TT
> clause if both are present; but
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> is laxer and allows either order.
</P
><P
> When using <TT
CLASS="LITERAL"
>LIMIT</TT
>, it is a good idea to use an
<TT
CLASS="LITERAL"
>ORDER BY</TT
> clause that constrains the result rows into a
unique order. Otherwise you will get an unpredictable subset of
the query's rows — you might be asking for the tenth through
twentieth rows, but tenth through twentieth in what ordering? You
don't know what ordering unless you specify <TT
CLASS="LITERAL"
>ORDER BY</TT
>.
</P
><P
> The query planner takes <TT
CLASS="LITERAL"
>LIMIT</TT
> into account when
generating a query plan, so you are very likely to get different
plans (yielding different row orders) depending on what you use
for <TT
CLASS="LITERAL"
>LIMIT</TT
> and <TT
CLASS="LITERAL"
>OFFSET</TT
>. Thus, using
different <TT
CLASS="LITERAL"
>LIMIT</TT
>/<TT
CLASS="LITERAL"
>OFFSET</TT
> values to select
different subsets of a query result <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>will give
inconsistent results</I
></SPAN
> unless you enforce a predictable
result ordering with <TT
CLASS="LITERAL"
>ORDER BY</TT
>. This is not a bug; it
is an inherent consequence of the fact that SQL does not promise
to deliver the results of a query in any particular order unless
<TT
CLASS="LITERAL"
>ORDER BY</TT
> is used to constrain the order.
</P
><P
> It is even possible for repeated executions of the same <TT
CLASS="LITERAL"
>LIMIT</TT
>
query to return different subsets of the rows of a table, if there
is not an <TT
CLASS="LITERAL"
>ORDER BY</TT
> to enforce selection of a deterministic
subset. Again, this is not a bug; determinism of the results is
simply not guaranteed in such a case.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-FOR-UPDATE-SHARE"
></A
><H3
><TT
CLASS="LITERAL"
>FOR UPDATE</TT
>/<TT
CLASS="LITERAL"
>FOR SHARE</TT
> Clause</H3
><P
> The <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> clause has this form:
</P><PRE
CLASS="SYNOPSIS"
>FOR UPDATE [ OF <TT
CLASS="REPLACEABLE"
><I
>table_name</I
></TT
> [, ...] ] [ NOWAIT ]</PRE
><P>
</P
><P
> The closely related <TT
CLASS="LITERAL"
>FOR SHARE</TT
> clause has this form:
</P><PRE
CLASS="SYNOPSIS"
>FOR SHARE [ OF <TT
CLASS="REPLACEABLE"
><I
>table_name</I
></TT
> [, ...] ] [ NOWAIT ]</PRE
><P>
</P
><P
> <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> causes the rows retrieved by the
<TT
CLASS="COMMAND"
>SELECT</TT
> statement to be locked as though for
update. This prevents them from being modified or deleted by
other transactions until the current transaction ends. That is,
other transactions that attempt <TT
CLASS="COMMAND"
>UPDATE</TT
>,
<TT
CLASS="COMMAND"
>DELETE</TT
>, or <TT
CLASS="COMMAND"
>SELECT FOR UPDATE</TT
>
of these rows will be blocked until the current transaction ends.
Also, if an <TT
CLASS="COMMAND"
>UPDATE</TT
>, <TT
CLASS="COMMAND"
>DELETE</TT
>,
or <TT
CLASS="COMMAND"
>SELECT FOR UPDATE</TT
> from another transaction
has already locked a selected row or rows, <TT
CLASS="COMMAND"
>SELECT FOR
UPDATE</TT
> will wait for the other transaction to complete,
and will then lock and return the updated row (or no row, if the
row was deleted). Within a <TT
CLASS="LITERAL"
>REPEATABLE READ</TT
> or <TT
CLASS="LITERAL"
>SERIALIZABLE</TT
> transaction,
however, an error will be thrown if a row to be locked has changed
since the transaction started. For further discussion see <A
HREF="mvcc.html"
>Chapter 13</A
>.
</P
><P
> <TT
CLASS="LITERAL"
>FOR SHARE</TT
> behaves similarly, except that it
acquires a shared rather than exclusive lock on each retrieved
row. A shared lock blocks other transactions from performing
<TT
CLASS="COMMAND"
>UPDATE</TT
>, <TT
CLASS="COMMAND"
>DELETE</TT
>, or <TT
CLASS="COMMAND"
>SELECT
FOR UPDATE</TT
> on these rows, but it does not prevent them
from performing <TT
CLASS="COMMAND"
>SELECT FOR SHARE</TT
>.
</P
><P
> To prevent the operation from waiting for other transactions to commit,
use the <TT
CLASS="LITERAL"
>NOWAIT</TT
> option. With <TT
CLASS="LITERAL"
>NOWAIT</TT
>, the statement
reports an error, rather than waiting, if a selected row
cannot be locked immediately. Note that <TT
CLASS="LITERAL"
>NOWAIT</TT
> applies only
to the row-level lock(s) — the required <TT
CLASS="LITERAL"
>ROW SHARE</TT
>
table-level lock is still taken in the ordinary way (see
<A
HREF="mvcc.html"
>Chapter 13</A
>). You can use
<A
HREF="sql-lock.html"
>LOCK</A
>
with the <TT
CLASS="LITERAL"
>NOWAIT</TT
> option first,
if you need to acquire the table-level lock without waiting.
</P
><P
> If specific tables are named in <TT
CLASS="LITERAL"
>FOR UPDATE</TT
>
or <TT
CLASS="LITERAL"
>FOR SHARE</TT
>,
then only rows coming from those tables are locked; any other
tables used in the <TT
CLASS="COMMAND"
>SELECT</TT
> are simply read as
usual. A <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> or <TT
CLASS="LITERAL"
>FOR SHARE</TT
>
clause without a table list affects all tables used in the statement.
If <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> or <TT
CLASS="LITERAL"
>FOR SHARE</TT
> is
applied to a view or sub-query, it affects all tables used in
the view or sub-query.
However, <TT
CLASS="LITERAL"
>FOR UPDATE</TT
>/<TT
CLASS="LITERAL"
>FOR SHARE</TT
>
do not apply to <TT
CLASS="LITERAL"
>WITH</TT
> queries referenced by the primary query.
If you want row locking to occur within a <TT
CLASS="LITERAL"
>WITH</TT
> query, specify
<TT
CLASS="LITERAL"
>FOR UPDATE</TT
> or <TT
CLASS="LITERAL"
>FOR SHARE</TT
> within the
<TT
CLASS="LITERAL"
>WITH</TT
> query.
</P
><P
> Multiple <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> and <TT
CLASS="LITERAL"
>FOR SHARE</TT
>
clauses can be written if it is necessary to specify different locking
behavior for different tables. If the same table is mentioned (or
implicitly affected) by both <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> and
<TT
CLASS="LITERAL"
>FOR SHARE</TT
> clauses, then it is processed as
<TT
CLASS="LITERAL"
>FOR UPDATE</TT
>. Similarly, a table is processed
as <TT
CLASS="LITERAL"
>NOWAIT</TT
> if that is specified in any of the clauses
affecting it.
</P
><P
> <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> and <TT
CLASS="LITERAL"
>FOR SHARE</TT
> cannot be
used in contexts where returned rows cannot be clearly identified with
individual table rows; for example they cannot be used with aggregation.
</P
><P
> When <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> or <TT
CLASS="LITERAL"
>FOR SHARE</TT
>
appears at the top level of a <TT
CLASS="COMMAND"
>SELECT</TT
> query, the rows that
are locked are exactly those that are returned by the query; in the
case of a join query, the rows locked are those that contribute to
returned join rows. In addition, rows that satisfied the query
conditions as of the query snapshot will be locked, although they
will not be returned if they were updated after the snapshot
and no longer satisfy the query conditions. If a
<TT
CLASS="LITERAL"
>LIMIT</TT
> is used, locking stops
once enough rows have been returned to satisfy the limit (but note that
rows skipped over by <TT
CLASS="LITERAL"
>OFFSET</TT
> will get locked). Similarly,
if <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> or <TT
CLASS="LITERAL"
>FOR SHARE</TT
>
is used in a cursor's query, only rows actually fetched or stepped past
by the cursor will be locked.
</P
><P
> When <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> or <TT
CLASS="LITERAL"
>FOR SHARE</TT
>
appears in a sub-<TT
CLASS="COMMAND"
>SELECT</TT
>, the rows locked are those
returned to the outer query by the sub-query. This might involve
fewer rows than inspection of the sub-query alone would suggest,
since conditions from the outer query might be used to optimize
execution of the sub-query. For example,
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT * FROM (SELECT * FROM mytable FOR UPDATE) ss WHERE col1 = 5;</PRE
><P>
will lock only rows having <TT
CLASS="LITERAL"
>col1 = 5</TT
>, even though that
condition is not textually within the sub-query.
</P
><DIV
CLASS="CAUTION"
><P
></P
><TABLE
CLASS="CAUTION"
BORDER="1"
WIDTH="100%"
><TR
><TD
ALIGN="CENTER"
><B
>Caution</B
></TD
></TR
><TR
><TD
ALIGN="LEFT"
><P
> Avoid locking a row and then modifying it within a later savepoint or
<SPAN
CLASS="APPLICATION"
>PL/pgSQL</SPAN
> exception block. A subsequent
rollback would cause the lock to be lost. For example:
</P><PRE
CLASS="PROGRAMLISTING"
>BEGIN;
SELECT * FROM mytable WHERE key = 1 FOR UPDATE;
SAVEPOINT s;
UPDATE mytable SET ... WHERE key = 1;
ROLLBACK TO s;</PRE
><P>
After the <TT
CLASS="COMMAND"
>ROLLBACK</TT
>, the row is effectively unlocked, rather
than returned to its pre-savepoint state of being locked but not modified.
This hazard occurs if a row locked in the current transaction is updated
or deleted, or if a shared lock is upgraded to exclusive: in all these
cases, the former lock state is forgotten. If the transaction is then
rolled back to a state between the original locking command and the
subsequent change, the row will appear not to be locked at all. This is
an implementation deficiency which will be addressed in a future release
of <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>.
</P
></TD
></TR
></TABLE
></DIV
><DIV
CLASS="CAUTION"
><P
></P
><TABLE
CLASS="CAUTION"
BORDER="1"
WIDTH="100%"
><TR
><TD
ALIGN="CENTER"
><B
>Caution</B
></TD
></TR
><TR
><TD
ALIGN="LEFT"
><P
> It is possible for a <TT
CLASS="COMMAND"
>SELECT</TT
> command running at the <TT
CLASS="LITERAL"
>READ
COMMITTED</TT
> transaction isolation level and using <TT
CLASS="LITERAL"
>ORDER
BY</TT
> and <TT
CLASS="LITERAL"
>FOR UPDATE/SHARE</TT
> to return rows out of
order. This is because <TT
CLASS="LITERAL"
>ORDER BY</TT
> is applied first.
The command sorts the result, but might then block trying to obtain a lock
on one or more of the rows. Once the <TT
CLASS="LITERAL"
>SELECT</TT
> unblocks, some
of the ordering column values might have been modified, leading to those
rows appearing to be out of order (though they are in order in terms
of the original column values). This can be worked around at need by
placing the <TT
CLASS="LITERAL"
>FOR UPDATE/SHARE</TT
> clause in a sub-query,
for example
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT * FROM (SELECT * FROM mytable FOR UPDATE) ss ORDER BY column1;</PRE
><P>
Note that this will result in locking all rows of <TT
CLASS="STRUCTNAME"
>mytable</TT
>,
whereas <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> at the top level would lock only the
actually returned rows. This can make for a significant performance
difference, particularly if the <TT
CLASS="LITERAL"
>ORDER BY</TT
> is combined with
<TT
CLASS="LITERAL"
>LIMIT</TT
> or other restrictions. So this technique is recommended
only if concurrent updates of the ordering columns are expected and a
strictly sorted result is required.
</P
><P
> At the <TT
CLASS="LITERAL"
>REPEATABLE READ</TT
> or <TT
CLASS="LITERAL"
>SERIALIZABLE</TT
>
transaction isolation level this would cause a serialization failure (with
a <TT
CLASS="LITERAL"
>SQLSTATE</TT
> of <TT
CLASS="LITERAL"
>'40001'</TT
>), so there is
no possibility of receiving rows out of order under these isolation levels.
</P
></TD
></TR
></TABLE
></DIV
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="SQL-TABLE"
></A
><H3
><TT
CLASS="LITERAL"
>TABLE</TT
> Command</H3
><P
> The command
</P><PRE
CLASS="PROGRAMLISTING"
>TABLE <TT
CLASS="REPLACEABLE"
><I
>name</I
></TT
></PRE
><P>
is completely equivalent to
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT * FROM <TT
CLASS="REPLACEABLE"
><I
>name</I
></TT
></PRE
><P>
It can be used as a top-level command or as a space-saving syntax
variant in parts of complex queries.
</P
></DIV
></DIV
><DIV
CLASS="REFSECT1"
><A
NAME="AEN78429"
></A
><H2
>Examples</H2
><P
> To join the table <TT
CLASS="LITERAL"
>films</TT
> with the table
<TT
CLASS="LITERAL"
>distributors</TT
>:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT f.title, f.did, d.name, f.date_prod, f.kind
FROM distributors d, films f
WHERE f.did = d.did
title | did | name | date_prod | kind
-------------------+-----+--------------+------------+----------
The Third Man | 101 | British Lion | 1949-12-23 | Drama
The African Queen | 101 | British Lion | 1951-08-11 | Romantic
...</PRE
><P>
</P
><P
> To sum the column <TT
CLASS="LITERAL"
>len</TT
> of all films and group
the results by <TT
CLASS="LITERAL"
>kind</TT
>:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT kind, sum(len) AS total FROM films GROUP BY kind;
kind | total
----------+-------
Action | 07:34
Comedy | 02:58
Drama | 14:28
Musical | 06:42
Romantic | 04:38</PRE
><P>
</P
><P
> To sum the column <TT
CLASS="LITERAL"
>len</TT
> of all films, group
the results by <TT
CLASS="LITERAL"
>kind</TT
> and show those group totals
that are less than 5 hours:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT kind, sum(len) AS total
FROM films
GROUP BY kind
HAVING sum(len) < interval '5 hours';
kind | total
----------+-------
Comedy | 02:58
Romantic | 04:38</PRE
><P>
</P
><P
> The following two examples are identical ways of sorting the individual
results according to the contents of the second column
(<TT
CLASS="LITERAL"
>name</TT
>):
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT * FROM distributors ORDER BY name;
SELECT * FROM distributors ORDER BY 2;
did | name
-----+------------------
109 | 20th Century Fox
110 | Bavaria Atelier
101 | British Lion
107 | Columbia
102 | Jean Luc Godard
113 | Luso films
104 | Mosfilm
103 | Paramount
106 | Toho
105 | United Artists
111 | Walt Disney
112 | Warner Bros.
108 | Westward</PRE
><P>
</P
><P
> The next example shows how to obtain the union of the tables
<TT
CLASS="LITERAL"
>distributors</TT
> and
<TT
CLASS="LITERAL"
>actors</TT
>, restricting the results to those that begin
with the letter W in each table. Only distinct rows are wanted, so the
key word <TT
CLASS="LITERAL"
>ALL</TT
> is omitted.
</P><PRE
CLASS="PROGRAMLISTING"
>distributors: actors:
did | name id | name
-----+-------------- ----+----------------
108 | Westward 1 | Woody Allen
111 | Walt Disney 2 | Warren Beatty
112 | Warner Bros. 3 | Walter Matthau
... ...
SELECT distributors.name
FROM distributors
WHERE distributors.name LIKE 'W%'
UNION
SELECT actors.name
FROM actors
WHERE actors.name LIKE 'W%';
name
----------------
Walt Disney
Walter Matthau
Warner Bros.
Warren Beatty
Westward
Woody Allen</PRE
><P>
</P
><P
> This example shows how to use a function in the <TT
CLASS="LITERAL"
>FROM</TT
>
clause, both with and without a column definition list:
</P><PRE
CLASS="PROGRAMLISTING"
>CREATE FUNCTION distributors(int) RETURNS SETOF distributors AS $$
SELECT * FROM distributors WHERE did = $1;
$$ LANGUAGE SQL;
SELECT * FROM distributors(111);
did | name
-----+-------------
111 | Walt Disney
CREATE FUNCTION distributors_2(int) RETURNS SETOF record AS $$
SELECT * FROM distributors WHERE did = $1;
$$ LANGUAGE SQL;
SELECT * FROM distributors_2(111) AS (f1 int, f2 text);
f1 | f2
-----+-------------
111 | Walt Disney</PRE
><P>
</P
><P
> This example shows how to use a simple <TT
CLASS="LITERAL"
>WITH</TT
> clause:
</P><PRE
CLASS="PROGRAMLISTING"
>WITH t AS (
SELECT random() as x FROM generate_series(1, 3)
)
SELECT * FROM t
UNION ALL
SELECT * FROM t
x
--------------------
0.534150459803641
0.520092216785997
0.0735620250925422
0.534150459803641
0.520092216785997
0.0735620250925422</PRE
><P>
Notice that the <TT
CLASS="LITERAL"
>WITH</TT
> query was evaluated only once,
so that we got two sets of the same three random values.
</P
><P
> This example uses <TT
CLASS="LITERAL"
>WITH RECURSIVE</TT
> to find all
subordinates (direct or indirect) of the employee Mary, and their
level of indirectness, from a table that shows only direct
subordinates:
</P><PRE
CLASS="PROGRAMLISTING"
>WITH RECURSIVE employee_recursive(distance, employee_name, manager_name) AS (
SELECT 1, employee_name, manager_name
FROM employee
WHERE manager_name = 'Mary'
UNION ALL
SELECT er.distance + 1, e.employee_name, e.manager_name
FROM employee_recursive er, employee e
WHERE er.employee_name = e.manager_name
)
SELECT distance, employee_name FROM employee_recursive;</PRE
><P>
Notice the typical form of recursive queries:
an initial condition, followed by <TT
CLASS="LITERAL"
>UNION</TT
>,
followed by the recursive part of the query. Be sure that the
recursive part of the query will eventually return no tuples, or
else the query will loop indefinitely. (See <A
HREF="queries-with.html"
>Section 7.8</A
>
for more examples.)
</P
></DIV
><DIV
CLASS="REFSECT1"
><A
NAME="AEN78463"
></A
><H2
>Compatibility</H2
><P
> Of course, the <TT
CLASS="COMMAND"
>SELECT</TT
> statement is compatible
with the SQL standard. But there are some extensions and some
missing features.
</P
><DIV
CLASS="REFSECT2"
><A
NAME="AEN78467"
></A
><H3
>Omitted <TT
CLASS="LITERAL"
>FROM</TT
> Clauses</H3
><P
> <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> allows one to omit the
<TT
CLASS="LITERAL"
>FROM</TT
> clause. It has a straightforward use to
compute the results of simple expressions:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT 2+2;
?column?
----------
4</PRE
><P>
Some other <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
> databases cannot do this except
by introducing a dummy one-row table from which to do the
<TT
CLASS="COMMAND"
>SELECT</TT
>.
</P
><P
> Note that if a <TT
CLASS="LITERAL"
>FROM</TT
> clause is not specified,
the query cannot reference any database tables. For example, the
following query is invalid:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT distributors.* WHERE distributors.name = 'Westward';</PRE
><P><SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> releases prior to
8.1 would accept queries of this form, and add an implicit entry
to the query's <TT
CLASS="LITERAL"
>FROM</TT
> clause for each table
referenced by the query. This is no longer allowed.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="AEN78481"
></A
><H3
>Omitting the <TT
CLASS="LITERAL"
>AS</TT
> Key Word</H3
><P
> In the SQL standard, the optional key word <TT
CLASS="LITERAL"
>AS</TT
> can be
omitted before an output column name whenever the new column name
is a valid column name (that is, not the same as any reserved
keyword). <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> is slightly more
restrictive: <TT
CLASS="LITERAL"
>AS</TT
> is required if the new column name
matches any keyword at all, reserved or not. Recommended practice is
to use <TT
CLASS="LITERAL"
>AS</TT
> or double-quote output column names, to prevent
any possible conflict against future keyword additions.
</P
><P
> In <TT
CLASS="LITERAL"
>FROM</TT
> items, both the standard and
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> allow <TT
CLASS="LITERAL"
>AS</TT
> to
be omitted before an alias that is an unreserved keyword. But
this is impractical for output column names, because of syntactic
ambiguities.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="AEN78493"
></A
><H3
><TT
CLASS="LITERAL"
>ONLY</TT
> and Inheritance</H3
><P
> The SQL standard requires parentheses around the table name when
writing <TT
CLASS="LITERAL"
>ONLY</TT
>, for example <TT
CLASS="LITERAL"
>SELECT * FROM ONLY
(tab1), ONLY (tab2) WHERE ...</TT
>. <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>
considers these parentheses to be optional.
</P
><P
> <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> allows a trailing <TT
CLASS="LITERAL"
>*</TT
> to be written to
explicitly specify the non-<TT
CLASS="LITERAL"
>ONLY</TT
> behavior of including
child tables. The standard does not allow this.
</P
><P
> (These points apply equally to all SQL commands supporting the
<TT
CLASS="LITERAL"
>ONLY</TT
> option.)
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="AEN78506"
></A
><H3
>Namespace Available to <TT
CLASS="LITERAL"
>GROUP BY</TT
> and <TT
CLASS="LITERAL"
>ORDER BY</TT
></H3
><P
> In the SQL-92 standard, an <TT
CLASS="LITERAL"
>ORDER BY</TT
> clause can
only use output column names or numbers, while a <TT
CLASS="LITERAL"
>GROUP
BY</TT
> clause can only use expressions based on input column
names. <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> extends each of
these clauses to allow the other choice as well (but it uses the
standard's interpretation if there is ambiguity).
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> also allows both clauses to
specify arbitrary expressions. Note that names appearing in an
expression will always be taken as input-column names, not as
output-column names.
</P
><P
> SQL:1999 and later use a slightly different definition which is not
entirely upward compatible with SQL-92.
In most cases, however, <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>
will interpret an <TT
CLASS="LITERAL"
>ORDER BY</TT
> or <TT
CLASS="LITERAL"
>GROUP
BY</TT
> expression the same way SQL:1999 does.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="AEN78519"
></A
><H3
>Functional Dependencies</H3
><P
> <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> recognizes functional dependency
(allowing columns to be omitted from <TT
CLASS="LITERAL"
>GROUP BY</TT
>) only when
a table's primary key is included in the <TT
CLASS="LITERAL"
>GROUP BY</TT
> list.
The SQL standard specifies additional conditions that should be
recognized.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="AEN78525"
></A
><H3
><TT
CLASS="LITERAL"
>WINDOW</TT
> Clause Restrictions</H3
><P
> The SQL standard provides additional options for the window
<TT
CLASS="REPLACEABLE"
><I
>frame_clause</I
></TT
>.
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> currently supports only the
options listed above.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="AEN78531"
></A
><H3
><TT
CLASS="LITERAL"
>LIMIT</TT
> and <TT
CLASS="LITERAL"
>OFFSET</TT
></H3
><P
> The clauses <TT
CLASS="LITERAL"
>LIMIT</TT
> and <TT
CLASS="LITERAL"
>OFFSET</TT
>
are <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>-specific syntax, also
used by <SPAN
CLASS="PRODUCTNAME"
>MySQL</SPAN
>. The SQL:2008 standard
has introduced the clauses <TT
CLASS="LITERAL"
>OFFSET ... FETCH {FIRST|NEXT}
...</TT
> for the same functionality, as shown above
in <A
HREF="sql-select.html#SQL-LIMIT"
><I
><I
>LIMIT</I
> Clause</I
></A
>. This
syntax is also used by <SPAN
CLASS="PRODUCTNAME"
>IBM DB2</SPAN
>.
(Applications written for <SPAN
CLASS="PRODUCTNAME"
>Oracle</SPAN
>
frequently use a workaround involving the automatically
generated <TT
CLASS="LITERAL"
>rownum</TT
> column, which is not available in
PostgreSQL, to implement the effects of these clauses.)
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="AEN78545"
></A
><H3
><TT
CLASS="LITERAL"
>FOR UPDATE</TT
> and <TT
CLASS="LITERAL"
>FOR SHARE</TT
></H3
><P
> Although <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> appears in the SQL standard, the
standard allows it only as an option of <TT
CLASS="COMMAND"
>DECLARE CURSOR</TT
>.
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> allows it in any <TT
CLASS="COMMAND"
>SELECT</TT
>
query as well as in sub-<TT
CLASS="COMMAND"
>SELECT</TT
>s, but this is an extension.
The <TT
CLASS="LITERAL"
>FOR SHARE</TT
> variant, and the <TT
CLASS="LITERAL"
>NOWAIT</TT
> option,
do not appear in the standard.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="AEN78557"
></A
><H3
>Data-Modifying Statements in <TT
CLASS="LITERAL"
>WITH</TT
></H3
><P
> <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> allows <TT
CLASS="COMMAND"
>INSERT</TT
>,
<TT
CLASS="COMMAND"
>UPDATE</TT
>, and <TT
CLASS="COMMAND"
>DELETE</TT
> to be used as <TT
CLASS="LITERAL"
>WITH</TT
>
queries. This is not found in the SQL standard.
</P
></DIV
><DIV
CLASS="REFSECT2"
><A
NAME="AEN78566"
></A
><H3
>Nonstandard Clauses</H3
><P
> The clause <TT
CLASS="LITERAL"
>DISTINCT ON</TT
> is not defined in the
SQL standard.
</P
></DIV
></DIV
><DIV
CLASS="NAVFOOTER"
><HR
ALIGN="LEFT"
WIDTH="100%"><TABLE
SUMMARY="Footer navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
><A
HREF="sql-security-label.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="index.html"
ACCESSKEY="H"
>Home</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
><A
HREF="sql-selectinto.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>SECURITY LABEL</TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="sql-commands.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
>SELECT INTO</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>