Current File : //opt/alt/postgresql11/usr/share/doc/alt-postgresql11-9.2.24/html/sql-expressions.html
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML
><HEAD
><TITLE
>Value Expressions</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 Syntax"
HREF="sql-syntax.html"><LINK
REL="PREVIOUS"
TITLE="Lexical Structure"
HREF="sql-syntax-lexical.html"><LINK
REL="NEXT"
TITLE="Calling Functions"
HREF="sql-syntax-calling-funcs.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="SECT1"
><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="Lexical Structure"
HREF="sql-syntax-lexical.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
HREF="sql-syntax.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="60%"
ALIGN="center"
VALIGN="bottom"
>Chapter 4. SQL Syntax</TD
><TD
WIDTH="20%"
ALIGN="right"
VALIGN="top"
><A
TITLE="Calling Functions"
HREF="sql-syntax-calling-funcs.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><DIV
CLASS="SECT1"
><H1
CLASS="SECT1"
><A
NAME="SQL-EXPRESSIONS"
>4.2. Value Expressions</A
></H1
><P
> Value expressions are used in a variety of contexts, such
as in the target list of the <TT
CLASS="COMMAND"
>SELECT</TT
> command, as
new column values in <TT
CLASS="COMMAND"
>INSERT</TT
> or
<TT
CLASS="COMMAND"
>UPDATE</TT
>, or in search conditions in a number of
commands. The result of a value expression is sometimes called a
<I
CLASS="FIRSTTERM"
>scalar</I
>, to distinguish it from the result of
a table expression (which is a table). Value expressions are
therefore also called <I
CLASS="FIRSTTERM"
>scalar expressions</I
> (or
even simply <I
CLASS="FIRSTTERM"
>expressions</I
>). The expression
syntax allows the calculation of values from primitive parts using
arithmetic, logical, set, and other operations.
</P
><P
> A value expression is one of the following:
<P
></P
></P><UL
><LI
><P
> A constant or literal value
</P
></LI
><LI
><P
> A column reference
</P
></LI
><LI
><P
> A positional parameter reference, in the body of a function definition
or prepared statement
</P
></LI
><LI
><P
> A subscripted expression
</P
></LI
><LI
><P
> A field selection expression
</P
></LI
><LI
><P
> An operator invocation
</P
></LI
><LI
><P
> A function call
</P
></LI
><LI
><P
> An aggregate expression
</P
></LI
><LI
><P
> A window function call
</P
></LI
><LI
><P
> A type cast
</P
></LI
><LI
><P
> A collation expression
</P
></LI
><LI
><P
> A scalar subquery
</P
></LI
><LI
><P
> An array constructor
</P
></LI
><LI
><P
> A row constructor
</P
></LI
><LI
><P
> Another value expression in parentheses (used to group
subexpressions and override
precedence)
</P
></LI
></UL
><P>
</P
><P
> In addition to this list, there are a number of constructs that can
be classified as an expression but do not follow any general syntax
rules. These generally have the semantics of a function or
operator and are explained in the appropriate location in <A
HREF="functions.html"
>Chapter 9</A
>. An example is the <TT
CLASS="LITERAL"
>IS NULL</TT
>
clause.
</P
><P
> We have already discussed constants in <A
HREF="sql-syntax-lexical.html#SQL-SYNTAX-CONSTANTS"
>Section 4.1.2</A
>. The following sections discuss
the remaining options.
</P
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SQL-EXPRESSIONS-COLUMN-REFS"
>4.2.1. Column References</A
></H2
><P
> A column can be referenced in the form:
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>correlation</I
></TT
>.<TT
CLASS="REPLACEABLE"
><I
>columnname</I
></TT
></PRE
><P>
</P
><P
> <TT
CLASS="REPLACEABLE"
><I
>correlation</I
></TT
> is the name of a
table (possibly qualified with a schema name), or an alias for a table
defined by means of a <TT
CLASS="LITERAL"
>FROM</TT
> clause.
The correlation name and separating dot can be omitted if the column name
is unique across all the tables being used in the current query. (See also <A
HREF="queries.html"
>Chapter 7</A
>.)
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SQL-EXPRESSIONS-PARAMETERS-POSITIONAL"
>4.2.2. Positional Parameters</A
></H2
><P
> A positional parameter reference is used to indicate a value
that is supplied externally to an SQL statement. Parameters are
used in SQL function definitions and in prepared queries. Some
client libraries also support specifying data values separately
from the SQL command string, in which case parameters are used to
refer to the out-of-line data values.
The form of a parameter reference is:
</P><PRE
CLASS="SYNOPSIS"
>$<TT
CLASS="REPLACEABLE"
><I
>number</I
></TT
></PRE
><P>
</P
><P
> For example, consider the definition of a function,
<CODE
CLASS="FUNCTION"
>dept</CODE
>, as:
</P><PRE
CLASS="PROGRAMLISTING"
>CREATE FUNCTION dept(text) RETURNS dept
AS $$ SELECT * FROM dept WHERE name = $1 $$
LANGUAGE SQL;</PRE
><P>
Here the <TT
CLASS="LITERAL"
>$1</TT
> references the value of the first
function argument whenever the function is invoked.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SQL-EXPRESSIONS-SUBSCRIPTS"
>4.2.3. Subscripts</A
></H2
><P
> If an expression yields a value of an array type, then a specific
element of the array value can be extracted by writing
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
>[<TT
CLASS="REPLACEABLE"
><I
>subscript</I
></TT
>]</PRE
><P>
or multiple adjacent elements (an <SPAN
CLASS="QUOTE"
>"array slice"</SPAN
>) can be extracted
by writing
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
>[<TT
CLASS="REPLACEABLE"
><I
>lower_subscript</I
></TT
>:<TT
CLASS="REPLACEABLE"
><I
>upper_subscript</I
></TT
>]</PRE
><P>
(Here, the brackets <TT
CLASS="LITERAL"
>[ ]</TT
> are meant to appear literally.)
Each <TT
CLASS="REPLACEABLE"
><I
>subscript</I
></TT
> is itself an expression,
which must yield an integer value.
</P
><P
> In general the array <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> must be
parenthesized, but the parentheses can be omitted when the expression
to be subscripted is just a column reference or positional parameter.
Also, multiple subscripts can be concatenated when the original array
is multidimensional.
For example:
</P><PRE
CLASS="PROGRAMLISTING"
>mytable.arraycolumn[4]
mytable.two_d_column[17][34]
$1[10:42]
(arrayfunction(a,b))[42]</PRE
><P>
The parentheses in the last example are required.
See <A
HREF="arrays.html"
>Section 8.15</A
> for more about arrays.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="FIELD-SELECTION"
>4.2.4. Field Selection</A
></H2
><P
> If an expression yields a value of a composite type (row type), then a
specific field of the row can be extracted by writing
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
>.<TT
CLASS="REPLACEABLE"
><I
>fieldname</I
></TT
></PRE
><P>
</P
><P
> In general the row <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> must be
parenthesized, but the parentheses can be omitted when the expression
to be selected from is just a table reference or positional parameter.
For example:
</P><PRE
CLASS="PROGRAMLISTING"
>mytable.mycolumn
$1.somecolumn
(rowfunction(a,b)).col3</PRE
><P>
(Thus, a qualified column reference is actually just a special case
of the field selection syntax.) An important special case is
extracting a field from a table column that is of a composite type:
</P><PRE
CLASS="PROGRAMLISTING"
>(compositecol).somefield
(mytable.compositecol).somefield</PRE
><P>
The parentheses are required here to show that
<TT
CLASS="STRUCTFIELD"
>compositecol</TT
> is a column name not a table name,
or that <TT
CLASS="STRUCTNAME"
>mytable</TT
> is a table name not a schema name
in the second case.
</P
><P
> You can ask for all fields of a composite value by
writing <TT
CLASS="LITERAL"
>.*</TT
>:
</P><PRE
CLASS="PROGRAMLISTING"
>(compositecol).*</PRE
><P>
This notation behaves differently depending on context;
see <A
HREF="rowtypes.html#ROWTYPES-USAGE"
>Section 8.16.5</A
> for details.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SQL-EXPRESSIONS-OPERATOR-CALLS"
>4.2.5. Operator Invocations</A
></H2
><P
> There are three possible syntaxes for an operator invocation:
<P
></P
><TABLE
BORDER="0"
><TBODY
><TR
><TD
><TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> <TT
CLASS="REPLACEABLE"
><I
>operator</I
></TT
> <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> (binary infix operator)</TD
></TR
><TR
><TD
><TT
CLASS="REPLACEABLE"
><I
>operator</I
></TT
> <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> (unary prefix operator)</TD
></TR
><TR
><TD
><TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> <TT
CLASS="REPLACEABLE"
><I
>operator</I
></TT
> (unary postfix operator)</TD
></TR
></TBODY
></TABLE
><P
></P
>
where the <TT
CLASS="REPLACEABLE"
><I
>operator</I
></TT
> token follows the syntax
rules of <A
HREF="sql-syntax-lexical.html#SQL-SYNTAX-OPERATORS"
>Section 4.1.3</A
>, or is one of the
key words <TT
CLASS="TOKEN"
>AND</TT
>, <TT
CLASS="TOKEN"
>OR</TT
>, and
<TT
CLASS="TOKEN"
>NOT</TT
>, or is a qualified operator name in the form:
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="LITERAL"
>OPERATOR(</TT
><TT
CLASS="REPLACEABLE"
><I
>schema</I
></TT
><TT
CLASS="LITERAL"
>.</TT
><TT
CLASS="REPLACEABLE"
><I
>operatorname</I
></TT
><TT
CLASS="LITERAL"
>)</TT
></PRE
><P>
Which particular operators exist and whether
they are unary or binary depends on what operators have been
defined by the system or the user. <A
HREF="functions.html"
>Chapter 9</A
>
describes the built-in operators.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SQL-EXPRESSIONS-FUNCTION-CALLS"
>4.2.6. Function Calls</A
></H2
><P
> The syntax for a function call is the name of a function
(possibly qualified with a schema name), followed by its argument list
enclosed in parentheses:
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>function_name</I
></TT
> ([<SPAN
CLASS="OPTIONAL"
><TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [<SPAN
CLASS="OPTIONAL"
>, <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> ... </SPAN
>]</SPAN
>] )</PRE
><P>
</P
><P
> For example, the following computes the square root of 2:
</P><PRE
CLASS="PROGRAMLISTING"
>sqrt(2)</PRE
><P>
</P
><P
> The list of built-in functions is in <A
HREF="functions.html"
>Chapter 9</A
>.
Other functions can be added by the user.
</P
><P
> The arguments can optionally have names attached.
See <A
HREF="sql-syntax-calling-funcs.html"
>Section 4.3</A
> for details.
</P
><DIV
CLASS="NOTE"
><BLOCKQUOTE
CLASS="NOTE"
><P
><B
>Note: </B
> A function that takes a single argument of composite type can
optionally be called using field-selection syntax, and conversely
field selection can be written in functional style. That is, the
notations <TT
CLASS="LITERAL"
>col(table)</TT
> and <TT
CLASS="LITERAL"
>table.col</TT
> are
interchangeable. This behavior is not SQL-standard but is provided
in <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> because it allows use of functions to
emulate <SPAN
CLASS="QUOTE"
>"computed fields"</SPAN
>. For more information see
<A
HREF="rowtypes.html#ROWTYPES-USAGE"
>Section 8.16.5</A
>.
</P
></BLOCKQUOTE
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SYNTAX-AGGREGATES"
>4.2.7. Aggregate Expressions</A
></H2
><P
> An <I
CLASS="FIRSTTERM"
>aggregate expression</I
> represents the
application of an aggregate function across the rows selected by a
query. An aggregate function reduces multiple inputs to a single
output value, such as the sum or average of the inputs. The
syntax of an aggregate expression is one of the following:
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>aggregate_name</I
></TT
> (<TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [ , ... ] [ <TT
CLASS="REPLACEABLE"
><I
>order_by_clause</I
></TT
> ] )
<TT
CLASS="REPLACEABLE"
><I
>aggregate_name</I
></TT
> (ALL <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [ , ... ] [ <TT
CLASS="REPLACEABLE"
><I
>order_by_clause</I
></TT
> ] )
<TT
CLASS="REPLACEABLE"
><I
>aggregate_name</I
></TT
> (DISTINCT <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [ , ... ] [ <TT
CLASS="REPLACEABLE"
><I
>order_by_clause</I
></TT
> ] )
<TT
CLASS="REPLACEABLE"
><I
>aggregate_name</I
></TT
> ( * )</PRE
><P>
where <TT
CLASS="REPLACEABLE"
><I
>aggregate_name</I
></TT
> is a previously
defined aggregate (possibly qualified with a schema name),
<TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> is
any value expression that does not itself contain an aggregate
expression or a window function call, and
<TT
CLASS="REPLACEABLE"
><I
>order_by_clause</I
></TT
> is a optional
<TT
CLASS="LITERAL"
>ORDER BY</TT
> clause as described below.
</P
><P
> The first form of aggregate expression invokes the aggregate
once for each input row.
The second form is the same as the first, since
<TT
CLASS="LITERAL"
>ALL</TT
> is the default.
The third form invokes the aggregate once for each distinct value
of the expression (or distinct set of values, for multiple expressions)
found in the input rows.
The last form invokes the aggregate once for each input row; since no
particular input value is specified, it is generally only useful
for the <CODE
CLASS="FUNCTION"
>count(*)</CODE
> aggregate function.
</P
><P
> Most aggregate functions ignore null inputs, so that rows in which
one or more of the expression(s) yield null are discarded. This
can be assumed to be true, unless otherwise specified, for all
built-in aggregates.
</P
><P
> For example, <TT
CLASS="LITERAL"
>count(*)</TT
> yields the total number
of input rows; <TT
CLASS="LITERAL"
>count(f1)</TT
> yields the number of
input rows in which <TT
CLASS="LITERAL"
>f1</TT
> is non-null, since
<CODE
CLASS="FUNCTION"
>count</CODE
> ignores nulls; and
<TT
CLASS="LITERAL"
>count(distinct f1)</TT
> yields the number of
distinct non-null values of <TT
CLASS="LITERAL"
>f1</TT
>.
</P
><P
> Ordinarily, the input rows are fed to the aggregate function in an
unspecified order. In many cases this does not matter; for example,
<CODE
CLASS="FUNCTION"
>min</CODE
> produces the same result no matter what order it
receives the inputs in. However, some aggregate functions
(such as <CODE
CLASS="FUNCTION"
>array_agg</CODE
> and <CODE
CLASS="FUNCTION"
>string_agg</CODE
>) produce
results that depend on the ordering of the input rows. When using
such an aggregate, the optional <TT
CLASS="REPLACEABLE"
><I
>order_by_clause</I
></TT
> can be
used to specify the desired ordering. The <TT
CLASS="REPLACEABLE"
><I
>order_by_clause</I
></TT
>
has the same syntax as for a query-level <TT
CLASS="LITERAL"
>ORDER BY</TT
> clause, as
described in <A
HREF="queries-order.html"
>Section 7.5</A
>, except that its expressions
are always just expressions and cannot be output-column names or numbers.
For example:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT array_agg(a ORDER BY b DESC) FROM table;</PRE
><P>
</P
><P
> When dealing with multiple-argument aggregate functions, note that the
<TT
CLASS="LITERAL"
>ORDER BY</TT
> clause goes after all the aggregate arguments.
For example, write this:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT string_agg(a, ',' ORDER BY a) FROM table;</PRE
><P>
not this:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT string_agg(a ORDER BY a, ',') FROM table; -- incorrect</PRE
><P>
The latter is syntactically valid, but it represents a call of a
single-argument aggregate function with two <TT
CLASS="LITERAL"
>ORDER BY</TT
> keys
(the second one being rather useless since it's a constant).
</P
><P
> If <TT
CLASS="LITERAL"
>DISTINCT</TT
> is specified in addition to an
<TT
CLASS="REPLACEABLE"
><I
>order_by_clause</I
></TT
>, then all the <TT
CLASS="LITERAL"
>ORDER BY</TT
>
expressions must match regular arguments of the aggregate; that is,
you cannot sort on an expression that is not included in the
<TT
CLASS="LITERAL"
>DISTINCT</TT
> list.
</P
><DIV
CLASS="NOTE"
><BLOCKQUOTE
CLASS="NOTE"
><P
><B
>Note: </B
> The ability to specify both <TT
CLASS="LITERAL"
>DISTINCT</TT
> and <TT
CLASS="LITERAL"
>ORDER BY</TT
>
in an aggregate function is a <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> extension.
</P
></BLOCKQUOTE
></DIV
><P
> The predefined aggregate functions are described in <A
HREF="functions-aggregate.html"
>Section 9.20</A
>. Other aggregate functions can be added
by the user.
</P
><P
> An aggregate expression can only appear in the result list or
<TT
CLASS="LITERAL"
>HAVING</TT
> clause of a <TT
CLASS="COMMAND"
>SELECT</TT
> command.
It is forbidden in other clauses, such as <TT
CLASS="LITERAL"
>WHERE</TT
>,
because those clauses are logically evaluated before the results
of aggregates are formed.
</P
><P
> When an aggregate expression appears in a subquery (see
<A
HREF="sql-expressions.html#SQL-SYNTAX-SCALAR-SUBQUERIES"
>Section 4.2.11</A
> and
<A
HREF="functions-subquery.html"
>Section 9.22</A
>), the aggregate is normally
evaluated over the rows of the subquery. But an exception occurs
if the aggregate's arguments contain only outer-level variables:
the aggregate then belongs to the nearest such outer level, and is
evaluated over the rows of that query. The aggregate expression
as a whole is then an outer reference for the subquery it appears in,
and acts as a constant over any one evaluation of that subquery.
The restriction about
appearing only in the result list or <TT
CLASS="LITERAL"
>HAVING</TT
> clause
applies with respect to the query level that the aggregate belongs to.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SYNTAX-WINDOW-FUNCTIONS"
>4.2.8. Window Function Calls</A
></H2
><P
> A <I
CLASS="FIRSTTERM"
>window function call</I
> represents the application
of an aggregate-like function over some portion of the rows selected
by a query. Unlike regular aggregate function calls, this is not tied
to grouping of the selected rows into a single output row — each
row remains separate in the query output. However the window function
is able to scan all the rows that would be part of the current row's
group according to the grouping specification (<TT
CLASS="LITERAL"
>PARTITION BY</TT
>
list) of the window function call.
The syntax of a window function call is one of the following:
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>function_name</I
></TT
> ([<SPAN
CLASS="OPTIONAL"
><TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [<SPAN
CLASS="OPTIONAL"
>, <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> ... </SPAN
>]</SPAN
>]) OVER <TT
CLASS="REPLACEABLE"
><I
>window_name</I
></TT
>
<TT
CLASS="REPLACEABLE"
><I
>function_name</I
></TT
> ([<SPAN
CLASS="OPTIONAL"
><TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> [<SPAN
CLASS="OPTIONAL"
>, <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> ... </SPAN
>]</SPAN
>]) OVER ( <TT
CLASS="REPLACEABLE"
><I
>window_definition</I
></TT
> )
<TT
CLASS="REPLACEABLE"
><I
>function_name</I
></TT
> ( * ) OVER <TT
CLASS="REPLACEABLE"
><I
>window_name</I
></TT
>
<TT
CLASS="REPLACEABLE"
><I
>function_name</I
></TT
> ( * ) OVER ( <TT
CLASS="REPLACEABLE"
><I
>window_definition</I
></TT
> )</PRE
><P>
where <TT
CLASS="REPLACEABLE"
><I
>window_definition</I
></TT
>
has the syntax
</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>
and the optional <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>
</P
><P
> Here, <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> represents any value
expression that does not itself contain window function calls.
The <TT
CLASS="LITERAL"
>PARTITION BY</TT
> and <TT
CLASS="LITERAL"
>ORDER BY</TT
> lists have
essentially the same syntax and semantics as <TT
CLASS="LITERAL"
>GROUP BY</TT
>
and <TT
CLASS="LITERAL"
>ORDER BY</TT
> clauses of the whole query, except that their
expressions are always just expressions and cannot be output-column
names or numbers.
<TT
CLASS="REPLACEABLE"
><I
>window_name</I
></TT
> is a reference to a named window
specification defined in the query's <TT
CLASS="LITERAL"
>WINDOW</TT
> clause.
Alternatively, a full <TT
CLASS="REPLACEABLE"
><I
>window_definition</I
></TT
> can
be given within parentheses, using the same syntax as for defining a
named window in the <TT
CLASS="LITERAL"
>WINDOW</TT
> clause; see the
<A
HREF="sql-select.html"
>SELECT</A
> reference page for details. It's worth
pointing out that <TT
CLASS="LITERAL"
>OVER wname</TT
> is not exactly equivalent to
<TT
CLASS="LITERAL"
>OVER (wname)</TT
>; the latter implies copying and modifying the
window definition, and will be rejected if the referenced window
specification includes a frame clause.
</P
><P
> The <TT
CLASS="REPLACEABLE"
><I
>frame_clause</I
></TT
> specifies
the set of rows constituting the <I
CLASS="FIRSTTERM"
>window frame</I
>, for those
window functions that act on the frame instead of the whole partition.
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.
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
> The built-in window functions are described in <A
HREF="functions-window.html#FUNCTIONS-WINDOW-TABLE"
>Table 9-48</A
>. Other window functions can be added by
the user. Also, any built-in or user-defined aggregate function can be
used as a window function.
</P
><P
> The syntaxes using <TT
CLASS="LITERAL"
>*</TT
> are used for calling parameter-less
aggregate functions as window functions, for example
<TT
CLASS="LITERAL"
>count(*) OVER (PARTITION BY x ORDER BY y)</TT
>.
The asterisk (<TT
CLASS="LITERAL"
>*</TT
>) is customarily not used for non-aggregate window functions.
Aggregate window functions, unlike normal aggregate functions, do not
allow <TT
CLASS="LITERAL"
>DISTINCT</TT
> or <TT
CLASS="LITERAL"
>ORDER BY</TT
> to be used within the
function argument list.
</P
><P
> Window function calls are permitted only in the <TT
CLASS="LITERAL"
>SELECT</TT
>
list and the <TT
CLASS="LITERAL"
>ORDER BY</TT
> clause of the query.
</P
><P
> More information about window functions can be found in
<A
HREF="tutorial-window.html"
>Section 3.5</A
>,
<A
HREF="functions-window.html"
>Section 9.21</A
>,
<A
HREF="queries-table-expressions.html#QUERIES-WINDOW"
>Section 7.2.4</A
>.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SQL-SYNTAX-TYPE-CASTS"
>4.2.9. Type Casts</A
></H2
><P
> A type cast specifies a conversion from one data type to another.
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> accepts two equivalent syntaxes
for type casts:
</P><PRE
CLASS="SYNOPSIS"
>CAST ( <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> AS <TT
CLASS="REPLACEABLE"
><I
>type</I
></TT
> )
<TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
>::<TT
CLASS="REPLACEABLE"
><I
>type</I
></TT
></PRE
><P>
The <TT
CLASS="LITERAL"
>CAST</TT
> syntax conforms to SQL; the syntax with
<TT
CLASS="LITERAL"
>::</TT
> is historical <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>
usage.
</P
><P
> When a cast is applied to a value expression of a known type, it
represents a run-time type conversion. The cast will succeed only
if a suitable type conversion operation has been defined. Notice that this
is subtly different from the use of casts with constants, as shown in
<A
HREF="sql-syntax-lexical.html#SQL-SYNTAX-CONSTANTS-GENERIC"
>Section 4.1.2.7</A
>. A cast applied to an
unadorned string literal represents the initial assignment of a type
to a literal constant value, and so it will succeed for any type
(if the contents of the string literal are acceptable input syntax for the
data type).
</P
><P
> An explicit type cast can usually be omitted if there is no ambiguity as
to the type that a value expression must produce (for example, when it is
assigned to a table column); the system will automatically apply a
type cast in such cases. However, automatic casting is only done for
casts that are marked <SPAN
CLASS="QUOTE"
>"OK to apply implicitly"</SPAN
>
in the system catalogs. Other casts must be invoked with
explicit casting syntax. This restriction is intended to prevent
surprising conversions from being applied silently.
</P
><P
> It is also possible to specify a type cast using a function-like
syntax:
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>typename</I
></TT
> ( <TT
CLASS="REPLACEABLE"
><I
>expression</I
></TT
> )</PRE
><P>
However, this only works for types whose names are also valid as
function names. For example, <TT
CLASS="LITERAL"
>double precision</TT
>
cannot be used this way, but the equivalent <TT
CLASS="LITERAL"
>float8</TT
>
can. Also, the names <TT
CLASS="LITERAL"
>interval</TT
>, <TT
CLASS="LITERAL"
>time</TT
>, and
<TT
CLASS="LITERAL"
>timestamp</TT
> can only be used in this fashion if they are
double-quoted, because of syntactic conflicts. Therefore, the use of
the function-like cast syntax leads to inconsistencies and should
probably be avoided.
</P
><DIV
CLASS="NOTE"
><BLOCKQUOTE
CLASS="NOTE"
><P
><B
>Note: </B
> The function-like syntax is in fact just a function call. When
one of the two standard cast syntaxes is used to do a run-time
conversion, it will internally invoke a registered function to
perform the conversion. By convention, these conversion functions
have the same name as their output type, and thus the <SPAN
CLASS="QUOTE"
>"function-like
syntax"</SPAN
> is nothing more than a direct invocation of the underlying
conversion function. Obviously, this is not something that a portable
application should rely on. For further details see
<A
HREF="sql-createcast.html"
>CREATE CAST</A
>.
</P
></BLOCKQUOTE
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SQL-SYNTAX-COLLATE-EXPRS"
>4.2.10. Collation Expressions</A
></H2
><P
> The <TT
CLASS="LITERAL"
>COLLATE</TT
> clause overrides the collation of
an expression. It is appended to the expression it applies to:
</P><PRE
CLASS="SYNOPSIS"
><TT
CLASS="REPLACEABLE"
><I
>expr</I
></TT
> COLLATE <TT
CLASS="REPLACEABLE"
><I
>collation</I
></TT
></PRE
><P>
where <TT
CLASS="REPLACEABLE"
><I
>collation</I
></TT
> is a possibly
schema-qualified identifier. The <TT
CLASS="LITERAL"
>COLLATE</TT
>
clause binds tighter than operators; parentheses can be used when
necessary.
</P
><P
> If no collation is explicitly specified, the database system
either derives a collation from the columns involved in the
expression, or it defaults to the default collation of the
database if no column is involved in the expression.
</P
><P
> The two common uses of the <TT
CLASS="LITERAL"
>COLLATE</TT
> clause are
overriding the sort order in an <TT
CLASS="LITERAL"
>ORDER BY</TT
> clause, for
example:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT a, b, c FROM tbl WHERE ... ORDER BY a COLLATE "C";</PRE
><P>
and overriding the collation of a function or operator call that
has locale-sensitive results, for example:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT * FROM tbl WHERE a > 'foo' COLLATE "C";</PRE
><P>
Note that in the latter case the <TT
CLASS="LITERAL"
>COLLATE</TT
> clause is
attached to an input argument of the operator we wish to affect.
It doesn't matter which argument of the operator or function call the
<TT
CLASS="LITERAL"
>COLLATE</TT
> clause is attached to, because the collation that is
applied by the operator or function is derived by considering all
arguments, and an explicit <TT
CLASS="LITERAL"
>COLLATE</TT
> clause will override the
collations of all other arguments. (Attaching non-matching
<TT
CLASS="LITERAL"
>COLLATE</TT
> clauses to more than one argument, however, is an
error. For more details see <A
HREF="collation.html"
>Section 22.2</A
>.)
Thus, this gives the same result as the previous example:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT * FROM tbl WHERE a COLLATE "C" > 'foo';</PRE
><P>
But this is an error:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT * FROM tbl WHERE (a > 'foo') COLLATE "C";</PRE
><P>
because it attempts to apply a collation to the result of the
<TT
CLASS="LITERAL"
>></TT
> operator, which is of the non-collatable data type
<TT
CLASS="TYPE"
>boolean</TT
>.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SQL-SYNTAX-SCALAR-SUBQUERIES"
>4.2.11. Scalar Subqueries</A
></H2
><P
> A scalar subquery is an ordinary
<TT
CLASS="COMMAND"
>SELECT</TT
> query in parentheses that returns exactly one
row with one column. (See <A
HREF="queries.html"
>Chapter 7</A
> for information about writing queries.)
The <TT
CLASS="COMMAND"
>SELECT</TT
> query is executed
and the single returned value is used in the surrounding value expression.
It is an error to use a query that
returns more than one row or more than one column as a scalar subquery.
(But if, during a particular execution, the subquery returns no rows,
there is no error; the scalar result is taken to be null.)
The subquery can refer to variables from the surrounding query,
which will act as constants during any one evaluation of the subquery.
See also <A
HREF="functions-subquery.html"
>Section 9.22</A
> for other expressions involving subqueries.
</P
><P
> For example, the following finds the largest city population in each
state:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT name, (SELECT max(pop) FROM cities WHERE cities.state = states.name)
FROM states;</PRE
><P>
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SQL-SYNTAX-ARRAY-CONSTRUCTORS"
>4.2.12. Array Constructors</A
></H2
><P
> An array constructor is an expression that builds an
array value using values for its member elements. A simple array
constructor
consists of the key word <TT
CLASS="LITERAL"
>ARRAY</TT
>, a left square bracket
<TT
CLASS="LITERAL"
>[</TT
>, a list of expressions (separated by commas) for the
array element values, and finally a right square bracket <TT
CLASS="LITERAL"
>]</TT
>.
For example:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT ARRAY[1,2,3+4];
array
---------
{1,2,7}
(1 row)</PRE
><P>
By default,
the array element type is the common type of the member expressions,
determined using the same rules as for <TT
CLASS="LITERAL"
>UNION</TT
> or
<TT
CLASS="LITERAL"
>CASE</TT
> constructs (see <A
HREF="typeconv-union-case.html"
>Section 10.5</A
>).
You can override this by explicitly casting the array constructor to the
desired type, for example:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT ARRAY[1,2,22.7]::integer[];
array
----------
{1,2,23}
(1 row)</PRE
><P>
This has the same effect as casting each expression to the array
element type individually.
For more on casting, see <A
HREF="sql-expressions.html#SQL-SYNTAX-TYPE-CASTS"
>Section 4.2.9</A
>.
</P
><P
> Multidimensional array values can be built by nesting array
constructors.
In the inner constructors, the key word <TT
CLASS="LITERAL"
>ARRAY</TT
> can
be omitted. For example, these produce the same result:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT ARRAY[ARRAY[1,2], ARRAY[3,4]];
array
---------------
{{1,2},{3,4}}
(1 row)
SELECT ARRAY[[1,2],[3,4]];
array
---------------
{{1,2},{3,4}}
(1 row)</PRE
><P>
Since multidimensional arrays must be rectangular, inner constructors
at the same level must produce sub-arrays of identical dimensions.
Any cast applied to the outer <TT
CLASS="LITERAL"
>ARRAY</TT
> constructor propagates
automatically to all the inner constructors.
</P
><P
> Multidimensional array constructor elements can be anything yielding
an array of the proper kind, not only a sub-<TT
CLASS="LITERAL"
>ARRAY</TT
> construct.
For example:
</P><PRE
CLASS="PROGRAMLISTING"
>CREATE TABLE arr(f1 int[], f2 int[]);
INSERT INTO arr VALUES (ARRAY[[1,2],[3,4]], ARRAY[[5,6],[7,8]]);
SELECT ARRAY[f1, f2, '{{9,10},{11,12}}'::int[]] FROM arr;
array
------------------------------------------------
{{{1,2},{3,4}},{{5,6},{7,8}},{{9,10},{11,12}}}
(1 row)</PRE
><P>
</P
><P
> You can construct an empty array, but since it's impossible to have an
array with no type, you must explicitly cast your empty array to the
desired type. For example:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT ARRAY[]::integer[];
array
-------
{}
(1 row)</PRE
><P>
</P
><P
> It is also possible to construct an array from the results of a
subquery. In this form, the array constructor is written with the
key word <TT
CLASS="LITERAL"
>ARRAY</TT
> followed by a parenthesized (not
bracketed) subquery. For example:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT ARRAY(SELECT oid FROM pg_proc WHERE proname LIKE 'bytea%');
array
-----------------------------------------------------------------------
{2011,1954,1948,1952,1951,1244,1950,2005,1949,1953,2006,31,2412,2413}
(1 row)</PRE
><P>
The subquery must return a single column. The resulting
one-dimensional array will have an element for each row in the
subquery result, with an element type matching that of the
subquery's output column.
</P
><P
> The subscripts of an array value built with <TT
CLASS="LITERAL"
>ARRAY</TT
>
always begin with one. For more information about arrays, see
<A
HREF="arrays.html"
>Section 8.15</A
>.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SQL-SYNTAX-ROW-CONSTRUCTORS"
>4.2.13. Row Constructors</A
></H2
><P
> A row constructor is an expression that builds a row value (also
called a composite value) using values
for its member fields. A row constructor consists of the key word
<TT
CLASS="LITERAL"
>ROW</TT
>, a left parenthesis, zero or more
expressions (separated by commas) for the row field values, and finally
a right parenthesis. For example:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT ROW(1,2.5,'this is a test');</PRE
><P>
The key word <TT
CLASS="LITERAL"
>ROW</TT
> is optional when there is more than one
expression in the list.
</P
><P
> A row constructor can include the syntax
<TT
CLASS="REPLACEABLE"
><I
>rowvalue</I
></TT
><TT
CLASS="LITERAL"
>.*</TT
>,
which will be expanded to a list of the elements of the row value,
just as occurs when the <TT
CLASS="LITERAL"
>.*</TT
> syntax is used at the top level
of a <TT
CLASS="COMMAND"
>SELECT</TT
> list (see <A
HREF="rowtypes.html#ROWTYPES-USAGE"
>Section 8.16.5</A
>).
For example, if table <TT
CLASS="LITERAL"
>t</TT
> has
columns <TT
CLASS="LITERAL"
>f1</TT
> and <TT
CLASS="LITERAL"
>f2</TT
>, these are the same:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT ROW(t.*, 42) FROM t;
SELECT ROW(t.f1, t.f2, 42) FROM t;</PRE
><P>
</P
><DIV
CLASS="NOTE"
><BLOCKQUOTE
CLASS="NOTE"
><P
><B
>Note: </B
> Before <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 8.2, the
<TT
CLASS="LITERAL"
>.*</TT
> syntax was not expanded in row constructors, so
that writing <TT
CLASS="LITERAL"
>ROW(t.*, 42)</TT
> created a two-field row whose first
field was another row value. The new behavior is usually more useful.
If you need the old behavior of nested row values, write the inner
row value without <TT
CLASS="LITERAL"
>.*</TT
>, for instance
<TT
CLASS="LITERAL"
>ROW(t, 42)</TT
>.
</P
></BLOCKQUOTE
></DIV
><P
> By default, the value created by a <TT
CLASS="LITERAL"
>ROW</TT
> expression is of
an anonymous record type. If necessary, it can be cast to a named
composite type — either the row type of a table, or a composite type
created with <TT
CLASS="COMMAND"
>CREATE TYPE AS</TT
>. An explicit cast might be needed
to avoid ambiguity. For example:
</P><PRE
CLASS="PROGRAMLISTING"
>CREATE TABLE mytable(f1 int, f2 float, f3 text);
CREATE FUNCTION getf1(mytable) RETURNS int AS 'SELECT $1.f1' LANGUAGE SQL;
-- No cast needed since only one getf1() exists
SELECT getf1(ROW(1,2.5,'this is a test'));
getf1
-------
1
(1 row)
CREATE TYPE myrowtype AS (f1 int, f2 text, f3 numeric);
CREATE FUNCTION getf1(myrowtype) RETURNS int AS 'SELECT $1.f1' LANGUAGE SQL;
-- Now we need a cast to indicate which function to call:
SELECT getf1(ROW(1,2.5,'this is a test'));
ERROR: function getf1(record) is not unique
SELECT getf1(ROW(1,2.5,'this is a test')::mytable);
getf1
-------
1
(1 row)
SELECT getf1(CAST(ROW(11,'this is a test',2.5) AS myrowtype));
getf1
-------
11
(1 row)</PRE
><P>
</P
><P
> Row constructors can be used to build composite values to be stored
in a composite-type table column, or to be passed to a function that
accepts a composite parameter. Also,
it is possible to compare two row values or test a row with
<TT
CLASS="LITERAL"
>IS NULL</TT
> or <TT
CLASS="LITERAL"
>IS NOT NULL</TT
>, for example:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT ROW(1,2.5,'this is a test') = ROW(1, 3, 'not the same');
SELECT ROW(table.*) IS NULL FROM table; -- detect all-null rows</PRE
><P>
For more detail see <A
HREF="functions-comparisons.html"
>Section 9.23</A
>.
Row constructors can also be used in connection with subqueries,
as discussed in <A
HREF="functions-subquery.html"
>Section 9.22</A
>.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="SYNTAX-EXPRESS-EVAL"
>4.2.14. Expression Evaluation Rules</A
></H2
><P
> The order of evaluation of subexpressions is not defined. In
particular, the inputs of an operator or function are not necessarily
evaluated left-to-right or in any other fixed order.
</P
><P
> Furthermore, if the result of an expression can be determined by
evaluating only some parts of it, then other subexpressions
might not be evaluated at all. For instance, if one wrote:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT true OR somefunc();</PRE
><P>
then <TT
CLASS="LITERAL"
>somefunc()</TT
> would (probably) not be called
at all. The same would be the case if one wrote:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT somefunc() OR true;</PRE
><P>
Note that this is not the same as the left-to-right
<SPAN
CLASS="QUOTE"
>"short-circuiting"</SPAN
> of Boolean operators that is found
in some programming languages.
</P
><P
> As a consequence, it is unwise to use functions with side effects
as part of complex expressions. It is particularly dangerous to
rely on side effects or evaluation order in <TT
CLASS="LITERAL"
>WHERE</TT
> and <TT
CLASS="LITERAL"
>HAVING</TT
> clauses,
since those clauses are extensively reprocessed as part of
developing an execution plan. Boolean
expressions (<TT
CLASS="LITERAL"
>AND</TT
>/<TT
CLASS="LITERAL"
>OR</TT
>/<TT
CLASS="LITERAL"
>NOT</TT
> combinations) in those clauses can be reorganized
in any manner allowed by the laws of Boolean algebra.
</P
><P
> When it is essential to force evaluation order, a <TT
CLASS="LITERAL"
>CASE</TT
>
construct (see <A
HREF="functions-conditional.html"
>Section 9.17</A
>) can be
used. For example, this is an untrustworthy way of trying to
avoid division by zero in a <TT
CLASS="LITERAL"
>WHERE</TT
> clause:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT ... WHERE x > 0 AND y/x > 1.5;</PRE
><P>
But this is safe:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT ... WHERE CASE WHEN x > 0 THEN y/x > 1.5 ELSE false END;</PRE
><P>
A <TT
CLASS="LITERAL"
>CASE</TT
> construct used in this fashion will defeat optimization
attempts, so it should only be done when necessary. (In this particular
example, it would be better to sidestep the problem by writing
<TT
CLASS="LITERAL"
>y > 1.5*x</TT
> instead.)
</P
><P
> <TT
CLASS="LITERAL"
>CASE</TT
> is not a cure-all for such issues, however.
One limitation of the technique illustrated above is that it does not
prevent early evaluation of constant subexpressions.
As described in <A
HREF="xfunc-volatility.html"
>Section 35.6</A
>, functions and
operators marked <TT
CLASS="LITERAL"
>IMMUTABLE</TT
> can be evaluated when
the query is planned rather than when it is executed. Thus for example
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT CASE WHEN x > 0 THEN x ELSE 1/0 END FROM tab;</PRE
><P>
is likely to result in a division-by-zero failure due to the planner
trying to simplify the constant subexpression,
even if every row in the table has <TT
CLASS="LITERAL"
>x > 0</TT
> so that the
<TT
CLASS="LITERAL"
>ELSE</TT
> arm would never be entered at run time.
</P
><P
> While that particular example might seem silly, related cases that don't
obviously involve constants can occur in queries executed within
functions, since the values of function arguments and local variables
can be inserted into queries as constants for planning purposes.
Within <SPAN
CLASS="APPLICATION"
>PL/pgSQL</SPAN
> functions, for example, using an
<TT
CLASS="LITERAL"
>IF</TT
>-<TT
CLASS="LITERAL"
>THEN</TT
>-<TT
CLASS="LITERAL"
>ELSE</TT
> statement to protect
a risky computation is much safer than just nesting it in a
<TT
CLASS="LITERAL"
>CASE</TT
> expression.
</P
><P
> Another limitation of the same kind is that a <TT
CLASS="LITERAL"
>CASE</TT
> cannot
prevent evaluation of an aggregate expression contained within it,
because aggregate expressions are computed before other
expressions in a <TT
CLASS="LITERAL"
>SELECT</TT
> list or <TT
CLASS="LITERAL"
>HAVING</TT
> clause
are considered. For example, the following query can cause a
division-by-zero error despite seemingly having protected against it:
</P><PRE
CLASS="PROGRAMLISTING"
>SELECT CASE WHEN min(employees) > 0
THEN avg(expenses / employees)
END
FROM departments;</PRE
><P>
The <CODE
CLASS="FUNCTION"
>min()</CODE
> and <CODE
CLASS="FUNCTION"
>avg()</CODE
> aggregates are computed
concurrently over all the input rows, so if any row
has <TT
CLASS="STRUCTFIELD"
>employees</TT
> equal to zero, the division-by-zero error
will occur before there is any opportunity to test the result of
<CODE
CLASS="FUNCTION"
>min()</CODE
>. Instead, use a <TT
CLASS="LITERAL"
>WHERE</TT
>
clause to prevent problematic input rows from
reaching an aggregate function in the first place.
</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-syntax-lexical.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-syntax-calling-funcs.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>Lexical Structure</TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="sql-syntax.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
>Calling Functions</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>
Mini Shell