<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>4.2. Value Expressions</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets Vsnapshot" /><link rel="prev" href="sql-syntax-lexical.html" title="4.1. Lexical Structure" /><link rel="next" href="sql-syntax-calling-funcs.html" title="4.3. Calling Functions" /></head><body><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">4.2. Value Expressions</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="sql-syntax-lexical.html" title="4.1. Lexical Structure">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="sql-syntax.html" title="Chapter 4. SQL Syntax">Up</a></td><th width="60%" align="center">Chapter 4. SQL Syntax</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 11.5 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="sql-syntax-calling-funcs.html" title="4.3. Calling Functions">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="SQL-EXPRESSIONS"><div class="titlepage"><div><div><h2 class="title" style="clear: both">4.2. Value Expressions</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="sql-expressions.html#SQL-EXPRESSIONS-COLUMN-REFS">4.2.1. Column References</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SQL-EXPRESSIONS-PARAMETERS-POSITIONAL">4.2.2. Positional Parameters</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SQL-EXPRESSIONS-SUBSCRIPTS">4.2.3. Subscripts</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#FIELD-SELECTION">4.2.4. Field Selection</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SQL-EXPRESSIONS-OPERATOR-CALLS">4.2.5. Operator Invocations</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SQL-EXPRESSIONS-FUNCTION-CALLS">4.2.6. Function Calls</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SYNTAX-AGGREGATES">4.2.7. Aggregate Expressions</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SYNTAX-WINDOW-FUNCTIONS">4.2.8. Window Function Calls</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SQL-SYNTAX-TYPE-CASTS">4.2.9. Type Casts</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SQL-SYNTAX-COLLATE-EXPRS">4.2.10. Collation Expressions</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SQL-SYNTAX-SCALAR-SUBQUERIES">4.2.11. Scalar Subqueries</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SQL-SYNTAX-ARRAY-CONSTRUCTORS">4.2.12. Array Constructors</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SQL-SYNTAX-ROW-CONSTRUCTORS">4.2.13. Row Constructors</a></span></dt><dt><span class="sect2"><a href="sql-expressions.html#SYNTAX-EXPRESS-EVAL">4.2.14. Expression Evaluation Rules</a></span></dt></dl></div><a id="id-1.5.3.6.2" class="indexterm"></a><a id="id-1.5.3.6.3" class="indexterm"></a><a id="id-1.5.3.6.4" class="indexterm"></a><p>
Value expressions are used in a variety of contexts, such
as in the target list of the <code class="command">SELECT</code> command, as
new column values in <code class="command">INSERT</code> or
<code class="command">UPDATE</code>, or in search conditions in a number of
commands. The result of a value expression is sometimes called a
<em class="firstterm">scalar</em>, to distinguish it from the result of
a table expression (which is a table). Value expressions are
therefore also called <em class="firstterm">scalar expressions</em> (or
even simply <em class="firstterm">expressions</em>). 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><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem"><p>
A constant or literal value
</p></li><li class="listitem"><p>
A column reference
</p></li><li class="listitem"><p>
A positional parameter reference, in the body of a function definition
or prepared statement
</p></li><li class="listitem"><p>
A subscripted expression
</p></li><li class="listitem"><p>
A field selection expression
</p></li><li class="listitem"><p>
An operator invocation
</p></li><li class="listitem"><p>
A function call
</p></li><li class="listitem"><p>
An aggregate expression
</p></li><li class="listitem"><p>
A window function call
</p></li><li class="listitem"><p>
A type cast
</p></li><li class="listitem"><p>
A collation expression
</p></li><li class="listitem"><p>
A scalar subquery
</p></li><li class="listitem"><p>
An array constructor
</p></li><li class="listitem"><p>
A row constructor
</p></li><li class="listitem"><p>
Another value expression in parentheses (used to group
subexpressions and override
precedence<a id="id-1.5.3.6.6.1.15.1.1" class="indexterm"></a>)
</p></li></ul></div><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 class="xref" href="functions.html" title="Chapter 9. Functions and Operators">Chapter 9</a>. An example is the <code class="literal">IS NULL</code>
clause.
</p><p>
We have already discussed constants in <a class="xref" href="sql-syntax-lexical.html#SQL-SYNTAX-CONSTANTS" title="4.1.2. Constants">Section 4.1.2</a>. The following sections discuss
the remaining options.
</p><div class="sect2" id="SQL-EXPRESSIONS-COLUMN-REFS"><div class="titlepage"><div><div><h3 class="title">4.2.1. Column References</h3></div></div></div><a id="id-1.5.3.6.9.2" class="indexterm"></a><p>
A column can be referenced in the form:
</p><pre class="synopsis">
<em class="replaceable"><code>correlation</code></em>.<em class="replaceable"><code>columnname</code></em>
</pre><p>
</p><p>
<em class="replaceable"><code>correlation</code></em> is the name of a
table (possibly qualified with a schema name), or an alias for a table
defined by means of a <code class="literal">FROM</code> 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 class="xref" href="queries.html" title="Chapter 7. Queries">Chapter 7</a>.)
</p></div><div class="sect2" id="SQL-EXPRESSIONS-PARAMETERS-POSITIONAL"><div class="titlepage"><div><div><h3 class="title">4.2.2. Positional Parameters</h3></div></div></div><a id="id-1.5.3.6.10.2" class="indexterm"></a><a id="id-1.5.3.6.10.3" class="indexterm"></a><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">
$<em class="replaceable"><code>number</code></em>
</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 <code class="literal">$1</code> references the value of the first
function argument whenever the function is invoked.
</p></div><div class="sect2" id="SQL-EXPRESSIONS-SUBSCRIPTS"><div class="titlepage"><div><div><h3 class="title">4.2.3. Subscripts</h3></div></div></div><a id="id-1.5.3.6.11.2" class="indexterm"></a><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">
<em class="replaceable"><code>expression</code></em>[<em class="replaceable"><code>subscript</code></em>]
</pre><p>
or multiple adjacent elements (an <span class="quote">“<span class="quote">array slice</span>”</span>) can be extracted
by writing
</p><pre class="synopsis">
<em class="replaceable"><code>expression</code></em>[<em class="replaceable"><code>lower_subscript</code></em>:<em class="replaceable"><code>upper_subscript</code></em>]
</pre><p>
(Here, the brackets <code class="literal">[ ]</code> are meant to appear literally.)
Each <em class="replaceable"><code>subscript</code></em> is itself an expression,
which must yield an integer value.
</p><p>
In general the array <em class="replaceable"><code>expression</code></em> 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 class="xref" href="arrays.html" title="8.15. Arrays">Section 8.15</a> for more about arrays.
</p></div><div class="sect2" id="FIELD-SELECTION"><div class="titlepage"><div><div><h3 class="title">4.2.4. Field Selection</h3></div></div></div><a id="id-1.5.3.6.12.2" class="indexterm"></a><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">
<em class="replaceable"><code>expression</code></em>.<em class="replaceable"><code>fieldname</code></em>
</pre><p>
</p><p>
In general the row <em class="replaceable"><code>expression</code></em> 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
<code class="structfield">compositecol</code> is a column name not a table name,
or that <code class="structname">mytable</code> 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 <code class="literal">.*</code>:
</p><pre class="programlisting">
(compositecol).*
</pre><p>
This notation behaves differently depending on context;
see <a class="xref" href="rowtypes.html#ROWTYPES-USAGE" title="8.16.5. Using Composite Types in Queries">Section 8.16.5</a> for details.
</p></div><div class="sect2" id="SQL-EXPRESSIONS-OPERATOR-CALLS"><div class="titlepage"><div><div><h3 class="title">4.2.5. Operator Invocations</h3></div></div></div><a id="id-1.5.3.6.13.2" class="indexterm"></a><p>
There are three possible syntaxes for an operator invocation:
</p><table border="0" summary="Simple list" class="simplelist"><tr><td><em class="replaceable"><code>expression</code></em> <em class="replaceable"><code>operator</code></em> <em class="replaceable"><code>expression</code></em> (binary infix operator)</td></tr><tr><td><em class="replaceable"><code>operator</code></em> <em class="replaceable"><code>expression</code></em> (unary prefix operator)</td></tr><tr><td><em class="replaceable"><code>expression</code></em> <em class="replaceable"><code>operator</code></em> (unary postfix operator)</td></tr></table><p>
where the <em class="replaceable"><code>operator</code></em> token follows the syntax
rules of <a class="xref" href="sql-syntax-lexical.html#SQL-SYNTAX-OPERATORS" title="4.1.3. Operators">Section 4.1.3</a>, or is one of the
key words <code class="token">AND</code>, <code class="token">OR</code>, and
<code class="token">NOT</code>, or is a qualified operator name in the form:
</p><pre class="synopsis">
<code class="literal">OPERATOR(</code><em class="replaceable"><code>schema</code></em><code class="literal">.</code><em class="replaceable"><code>operatorname</code></em><code class="literal">)</code>
</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 class="xref" href="functions.html" title="Chapter 9. Functions and Operators">Chapter 9</a>
describes the built-in operators.
</p></div><div class="sect2" id="SQL-EXPRESSIONS-FUNCTION-CALLS"><div class="titlepage"><div><div><h3 class="title">4.2.6. Function Calls</h3></div></div></div><a id="id-1.5.3.6.14.2" class="indexterm"></a><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">
<em class="replaceable"><code>function_name</code></em> ([<span class="optional"><em class="replaceable"><code>expression</code></em> [<span class="optional">, <em class="replaceable"><code>expression</code></em> ... </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 class="xref" href="functions.html" title="Chapter 9. Functions and Operators">Chapter 9</a>.
Other functions can be added by the user.
</p><p>
When issuing queries in a database where some users mistrust other users,
observe security precautions from <a class="xref" href="typeconv-func.html" title="10.3. Functions">Section 10.3</a> when
writing function calls.
</p><p>
The arguments can optionally have names attached.
See <a class="xref" href="sql-syntax-calling-funcs.html" title="4.3. Calling Functions">Section 4.3</a> for details.
</p><div class="note"><h3 class="title">Note</h3><p>
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 <code class="literal">col(table)</code> and <code class="literal">table.col</code> 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">“<span class="quote">computed fields</span>”</span>. For more information see
<a class="xref" href="rowtypes.html#ROWTYPES-USAGE" title="8.16.5. Using Composite Types in Queries">Section 8.16.5</a>.
</p></div></div><div class="sect2" id="SYNTAX-AGGREGATES"><div class="titlepage"><div><div><h3 class="title">4.2.7. Aggregate Expressions</h3></div></div></div><a id="id-1.5.3.6.15.2" class="indexterm"></a><a id="id-1.5.3.6.15.3" class="indexterm"></a><a id="id-1.5.3.6.15.4" class="indexterm"></a><a id="id-1.5.3.6.15.5" class="indexterm"></a><p>
An <em class="firstterm">aggregate expression</em> 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">
<em class="replaceable"><code>aggregate_name</code></em> (<em class="replaceable"><code>expression</code></em> [ , ... ] [ <em class="replaceable"><code>order_by_clause</code></em> ] ) [ FILTER ( WHERE <em class="replaceable"><code>filter_clause</code></em> ) ]
<em class="replaceable"><code>aggregate_name</code></em> (ALL <em class="replaceable"><code>expression</code></em> [ , ... ] [ <em class="replaceable"><code>order_by_clause</code></em> ] ) [ FILTER ( WHERE <em class="replaceable"><code>filter_clause</code></em> ) ]
<em class="replaceable"><code>aggregate_name</code></em> (DISTINCT <em class="replaceable"><code>expression</code></em> [ , ... ] [ <em class="replaceable"><code>order_by_clause</code></em> ] ) [ FILTER ( WHERE <em class="replaceable"><code>filter_clause</code></em> ) ]
<em class="replaceable"><code>aggregate_name</code></em> ( * ) [ FILTER ( WHERE <em class="replaceable"><code>filter_clause</code></em> ) ]
<em class="replaceable"><code>aggregate_name</code></em> ( [ <em class="replaceable"><code>expression</code></em> [ , ... ] ] ) WITHIN GROUP ( <em class="replaceable"><code>order_by_clause</code></em> ) [ FILTER ( WHERE <em class="replaceable"><code>filter_clause</code></em> ) ]
</pre><p>
where <em class="replaceable"><code>aggregate_name</code></em> is a previously
defined aggregate (possibly qualified with a schema name) and
<em class="replaceable"><code>expression</code></em> is
any value expression that does not itself contain an aggregate
expression or a window function call. The optional
<em class="replaceable"><code>order_by_clause</code></em> and
<em class="replaceable"><code>filter_clause</code></em> are 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
<code class="literal">ALL</code> 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 fourth 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.
The last form is used with <em class="firstterm">ordered-set</em> aggregate
functions, which are described below.
</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, <code class="literal">count(*)</code> yields the total number
of input rows; <code class="literal">count(f1)</code> yields the number of
input rows in which <code class="literal">f1</code> is non-null, since
<code class="function">count</code> ignores nulls; and
<code class="literal">count(distinct f1)</code> yields the number of
distinct non-null values of <code class="literal">f1</code>.
</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 <em class="replaceable"><code>order_by_clause</code></em> can be
used to specify the desired ordering. The <em class="replaceable"><code>order_by_clause</code></em>
has the same syntax as for a query-level <code class="literal">ORDER BY</code> clause, as
described in <a class="xref" href="queries-order.html" title="7.5. Sorting Rows">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
<code class="literal">ORDER BY</code> 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 <code class="literal">ORDER BY</code> keys
(the second one being rather useless since it's a constant).
</p><p>
If <code class="literal">DISTINCT</code> is specified in addition to an
<em class="replaceable"><code>order_by_clause</code></em>, then all the <code class="literal">ORDER BY</code>
expressions must match regular arguments of the aggregate; that is,
you cannot sort on an expression that is not included in the
<code class="literal">DISTINCT</code> list.
</p><div class="note"><h3 class="title">Note</h3><p>
The ability to specify both <code class="literal">DISTINCT</code> and <code class="literal">ORDER BY</code>
in an aggregate function is a <span class="productname">PostgreSQL</span> extension.
</p></div><p>
Placing <code class="literal">ORDER BY</code> within the aggregate's regular argument
list, as described so far, is used when ordering the input rows for
general-purpose and statistical aggregates, for which ordering is
optional. There is a
subclass of aggregate functions called <em class="firstterm">ordered-set
aggregates</em> for which an <em class="replaceable"><code>order_by_clause</code></em>
is <span class="emphasis"><em>required</em></span>, usually because the aggregate's computation is
only sensible in terms of a specific ordering of its input rows.
Typical examples of ordered-set aggregates include rank and percentile
calculations. For an ordered-set aggregate,
the <em class="replaceable"><code>order_by_clause</code></em> is written
inside <code class="literal">WITHIN GROUP (...)</code>, as shown in the final syntax
alternative above. The expressions in
the <em class="replaceable"><code>order_by_clause</code></em> are evaluated once per
input row just like regular aggregate arguments, sorted as per
the <em class="replaceable"><code>order_by_clause</code></em>'s requirements, and fed
to the aggregate function as input arguments. (This is unlike the case
for a non-<code class="literal">WITHIN GROUP</code> <em class="replaceable"><code>order_by_clause</code></em>,
which is not treated as argument(s) to the aggregate function.) The
argument expressions preceding <code class="literal">WITHIN GROUP</code>, if any, are
called <em class="firstterm">direct arguments</em> to distinguish them from
the <em class="firstterm">aggregated arguments</em> listed in
the <em class="replaceable"><code>order_by_clause</code></em>. Unlike regular aggregate
arguments, direct arguments are evaluated only once per aggregate call,
not once per input row. This means that they can contain variables only
if those variables are grouped by <code class="literal">GROUP BY</code>; this restriction
is the same as if the direct arguments were not inside an aggregate
expression at all. Direct arguments are typically used for things like
percentile fractions, which only make sense as a single value per
aggregation calculation. The direct argument list can be empty; in this
case, write just <code class="literal">()</code> not <code class="literal">(*)</code>.
(<span class="productname">PostgreSQL</span> will actually accept either spelling, but
only the first way conforms to the SQL standard.)
</p><p>
<a id="id-1.5.3.6.15.15.1" class="indexterm"></a>
An example of an ordered-set aggregate call is:
</p><pre class="programlisting">
SELECT percentile_cont(0.5) WITHIN GROUP (ORDER BY income) FROM households;
percentile_cont
-----------------
50489
</pre><p>
which obtains the 50th percentile, or median, value of
the <code class="structfield">income</code> column from table <code class="structname">households</code>.
Here, <code class="literal">0.5</code> is a direct argument; it would make no sense
for the percentile fraction to be a value varying across rows.
</p><p>
If <code class="literal">FILTER</code> is specified, then only the input
rows for which the <em class="replaceable"><code>filter_clause</code></em>
evaluates to true are fed to the aggregate function; other rows
are discarded. For example:
</p><pre class="programlisting">
SELECT
count(*) AS unfiltered,
count(*) FILTER (WHERE i < 5) AS filtered
FROM generate_series(1,10) AS s(i);
unfiltered | filtered
------------+----------
10 | 4
(1 row)
</pre><p>
</p><p>
The predefined aggregate functions are described in <a class="xref" href="functions-aggregate.html" title="9.20. Aggregate Functions">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
<code class="literal">HAVING</code> clause of a <code class="command">SELECT</code> command.
It is forbidden in other clauses, such as <code class="literal">WHERE</code>,
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 class="xref" href="sql-expressions.html#SQL-SYNTAX-SCALAR-SUBQUERIES" title="4.2.11. Scalar Subqueries">Section 4.2.11</a> and
<a class="xref" href="functions-subquery.html" title="9.22. Subquery Expressions">Section 9.22</a>), the aggregate is normally
evaluated over the rows of the subquery. But an exception occurs
if the aggregate's arguments (and <em class="replaceable"><code>filter_clause</code></em>
if any) 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 <code class="literal">HAVING</code> clause
applies with respect to the query level that the aggregate belongs to.
</p></div><div class="sect2" id="SYNTAX-WINDOW-FUNCTIONS"><div class="titlepage"><div><div><h3 class="title">4.2.8. Window Function Calls</h3></div></div></div><a id="id-1.5.3.6.16.2" class="indexterm"></a><a id="id-1.5.3.6.16.3" class="indexterm"></a><p>
A <em class="firstterm">window function call</em> represents the application
of an aggregate-like function over some portion of the rows selected
by a query. Unlike non-window aggregate 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
has access to all the rows that would be part of the current row's
group according to the grouping specification (<code class="literal">PARTITION BY</code>
list) of the window function call.
The syntax of a window function call is one of the following:
</p><pre class="synopsis">
<em class="replaceable"><code>function_name</code></em> ([<span class="optional"><em class="replaceable"><code>expression</code></em> [<span class="optional">, <em class="replaceable"><code>expression</code></em> ... </span>]</span>]) [ FILTER ( WHERE <em class="replaceable"><code>filter_clause</code></em> ) ] OVER <em class="replaceable"><code>window_name</code></em>
<em class="replaceable"><code>function_name</code></em> ([<span class="optional"><em class="replaceable"><code>expression</code></em> [<span class="optional">, <em class="replaceable"><code>expression</code></em> ... </span>]</span>]) [ FILTER ( WHERE <em class="replaceable"><code>filter_clause</code></em> ) ] OVER ( <em class="replaceable"><code>window_definition</code></em> )
<em class="replaceable"><code>function_name</code></em> ( * ) [ FILTER ( WHERE <em class="replaceable"><code>filter_clause</code></em> ) ] OVER <em class="replaceable"><code>window_name</code></em>
<em class="replaceable"><code>function_name</code></em> ( * ) [ FILTER ( WHERE <em class="replaceable"><code>filter_clause</code></em> ) ] OVER ( <em class="replaceable"><code>window_definition</code></em> )
</pre><p>
where <em class="replaceable"><code>window_definition</code></em>
has the syntax
</p><pre class="synopsis">
[ <em class="replaceable"><code>existing_window_name</code></em> ]
[ PARTITION BY <em class="replaceable"><code>expression</code></em> [, ...] ]
[ ORDER BY <em class="replaceable"><code>expression</code></em> [ ASC | DESC | USING <em class="replaceable"><code>operator</code></em> ] [ NULLS { FIRST | LAST } ] [, ...] ]
[ <em class="replaceable"><code>frame_clause</code></em> ]
</pre><p>
The optional <em class="replaceable"><code>frame_clause</code></em>
can be one of
</p><pre class="synopsis">
{ RANGE | ROWS | GROUPS } <em class="replaceable"><code>frame_start</code></em> [ <em class="replaceable"><code>frame_exclusion</code></em> ]
{ RANGE | ROWS | GROUPS } BETWEEN <em class="replaceable"><code>frame_start</code></em> AND <em class="replaceable"><code>frame_end</code></em> [ <em class="replaceable"><code>frame_exclusion</code></em> ]
</pre><p>
where <em class="replaceable"><code>frame_start</code></em>
and <em class="replaceable"><code>frame_end</code></em> can be one of
</p><pre class="synopsis">
UNBOUNDED PRECEDING
<em class="replaceable"><code>offset</code></em> PRECEDING
CURRENT ROW
<em class="replaceable"><code>offset</code></em> FOLLOWING
UNBOUNDED FOLLOWING
</pre><p>
and <em class="replaceable"><code>frame_exclusion</code></em> can be one of
</p><pre class="synopsis">
EXCLUDE CURRENT ROW
EXCLUDE GROUP
EXCLUDE TIES
EXCLUDE NO OTHERS
</pre><p>
</p><p>
Here, <em class="replaceable"><code>expression</code></em> represents any value
expression that does not itself contain window function calls.
</p><p>
<em class="replaceable"><code>window_name</code></em> is a reference to a named window
specification defined in the query's <code class="literal">WINDOW</code> clause.
Alternatively, a full <em class="replaceable"><code>window_definition</code></em> can
be given within parentheses, using the same syntax as for defining a
named window in the <code class="literal">WINDOW</code> clause; see the
<a class="xref" href="sql-select.html" title="SELECT"><span class="refentrytitle">SELECT</span></a> reference page for details. It's worth
pointing out that <code class="literal">OVER wname</code> is not exactly equivalent to
<code class="literal">OVER (wname ...)</code>; 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 <code class="literal">PARTITION BY</code> clause groups the rows of the query into
<em class="firstterm">partitions</em>, which are processed separately by the window
function. <code class="literal">PARTITION BY</code> works similarly to a query-level
<code class="literal">GROUP BY</code> clause, except that its expressions are always just
expressions and cannot be output-column names or numbers.
Without <code class="literal">PARTITION BY</code>, all rows produced by the query are
treated as a single partition.
The <code class="literal">ORDER BY</code> clause determines the order in which the rows
of a partition are processed by the window function. It works similarly
to a query-level <code class="literal">ORDER BY</code> clause, but likewise cannot use
output-column names or numbers. Without <code class="literal">ORDER BY</code>, rows are
processed in an unspecified order.
</p><p>
The <em class="replaceable"><code>frame_clause</code></em> specifies
the set of rows constituting the <em class="firstterm">window frame</em>, which is a
subset of the current partition, for those window functions that act on
the frame instead of the whole partition. The set of rows in the frame
can vary depending on which row is the current row. The frame can be
specified in <code class="literal">RANGE</code>, <code class="literal">ROWS</code>
or <code class="literal">GROUPS</code> mode; in each case, it runs from
the <em class="replaceable"><code>frame_start</code></em> to
the <em class="replaceable"><code>frame_end</code></em>.
If <em class="replaceable"><code>frame_end</code></em> is omitted, the end defaults
to <code class="literal">CURRENT ROW</code>.
</p><p>
A <em class="replaceable"><code>frame_start</code></em> of <code class="literal">UNBOUNDED PRECEDING</code> means
that the frame starts with the first row of the partition, and similarly
a <em class="replaceable"><code>frame_end</code></em> of <code class="literal">UNBOUNDED FOLLOWING</code> means
that the frame ends with the last row of the partition.
</p><p>
In <code class="literal">RANGE</code> or <code class="literal">GROUPS</code> mode,
a <em class="replaceable"><code>frame_start</code></em> of
<code class="literal">CURRENT ROW</code> means the frame starts with the current
row's first <em class="firstterm">peer</em> row (a row that the
window's <code class="literal">ORDER BY</code> clause sorts as equivalent to the
current row), while a <em class="replaceable"><code>frame_end</code></em> of
<code class="literal">CURRENT ROW</code> means the frame ends with the current
row's last peer row.
In <code class="literal">ROWS</code> mode, <code class="literal">CURRENT ROW</code> simply
means the current row.
</p><p>
In the <em class="replaceable"><code>offset</code></em> <code class="literal">PRECEDING</code>
and <em class="replaceable"><code>offset</code></em> <code class="literal">FOLLOWING</code> frame
options, the <em class="replaceable"><code>offset</code></em> must be an expression not
containing any variables, aggregate functions, or window functions.
The meaning of the <em class="replaceable"><code>offset</code></em> depends on the
frame mode:
</p><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem"><p>
In <code class="literal">ROWS</code> mode,
the <em class="replaceable"><code>offset</code></em> must yield a non-null,
non-negative integer, and the option means that the frame starts or
ends the specified number of rows before or after the current row.
</p></li><li class="listitem"><p>
In <code class="literal">GROUPS</code> mode,
the <em class="replaceable"><code>offset</code></em> again must yield a non-null,
non-negative integer, and the option means that the frame starts or
ends the specified number of <em class="firstterm">peer groups</em>
before or after the current row's peer group, where a peer group is a
set of rows that are equivalent in the <code class="literal">ORDER BY</code>
ordering. (There must be an <code class="literal">ORDER BY</code> clause
in the window definition to use <code class="literal">GROUPS</code> mode.)
</p></li><li class="listitem"><p>
In <code class="literal">RANGE</code> mode, these options require that
the <code class="literal">ORDER BY</code> clause specify exactly one column.
The <em class="replaceable"><code>offset</code></em> specifies the maximum
difference between the value of that column in the current row and
its value in preceding or following rows of the frame. The data type
of the <em class="replaceable"><code>offset</code></em> expression varies depending
on the data type of the ordering column. For numeric ordering
columns it is typically of the same type as the ordering column,
but for datetime ordering columns it is an <code class="type">interval</code>.
For example, if the ordering column is of type <code class="type">date</code>
or <code class="type">timestamp</code>, one could write <code class="literal">RANGE BETWEEN
'1 day' PRECEDING AND '10 days' FOLLOWING</code>.
The <em class="replaceable"><code>offset</code></em> is still required to be
non-null and non-negative, though the meaning
of <span class="quote">“<span class="quote">non-negative</span>”</span> depends on its data type.
</p></li></ul></div><p>
In any case, the distance to the end of the frame is limited by the
distance to the end of the partition, so that for rows near the partition
ends the frame might contain fewer rows than elsewhere.
</p><p>
Notice that in both <code class="literal">ROWS</code> and <code class="literal">GROUPS</code>
mode, <code class="literal">0 PRECEDING</code> and <code class="literal">0 FOLLOWING</code>
are equivalent to <code class="literal">CURRENT ROW</code>. This normally holds
in <code class="literal">RANGE</code> mode as well, for an appropriate
data-type-specific meaning of <span class="quote">“<span class="quote">zero</span>”</span>.
</p><p>
The <em class="replaceable"><code>frame_exclusion</code></em> option allows rows around
the current row to be excluded from the frame, even if they would be
included according to the frame start and frame end options.
<code class="literal">EXCLUDE CURRENT ROW</code> excludes the current row from the
frame.
<code class="literal">EXCLUDE GROUP</code> excludes the current row and its
ordering peers from the frame.
<code class="literal">EXCLUDE TIES</code> excludes any peers of the current
row from the frame, but not the current row itself.
<code class="literal">EXCLUDE NO OTHERS</code> simply specifies explicitly the
default behavior of not excluding the current row or its peers.
</p><p>
The default framing option is <code class="literal">RANGE UNBOUNDED PRECEDING</code>,
which is the same as <code class="literal">RANGE BETWEEN UNBOUNDED PRECEDING AND
CURRENT ROW</code>. With <code class="literal">ORDER BY</code>, this sets the frame to be
all rows from the partition start up through the current row's last
<code class="literal">ORDER BY</code> peer. Without <code class="literal">ORDER BY</code>,
this means all rows of the partition are included in the window frame,
since all rows become peers of the current row.
</p><p>
Restrictions are that
<em class="replaceable"><code>frame_start</code></em> cannot be <code class="literal">UNBOUNDED FOLLOWING</code>,
<em class="replaceable"><code>frame_end</code></em> cannot be <code class="literal">UNBOUNDED PRECEDING</code>,
and the <em class="replaceable"><code>frame_end</code></em> choice cannot appear earlier in the
above list of <em class="replaceable"><code>frame_start</code></em>
and <em class="replaceable"><code>frame_end</code></em> options than
the <em class="replaceable"><code>frame_start</code></em> choice does — for example
<code class="literal">RANGE BETWEEN CURRENT ROW AND <em class="replaceable"><code>offset</code></em>
PRECEDING</code> is not allowed.
But, for example, <code class="literal">ROWS BETWEEN 7 PRECEDING AND 8
PRECEDING</code> is allowed, even though it would never select any
rows.
</p><p>
If <code class="literal">FILTER</code> is specified, then only the input
rows for which the <em class="replaceable"><code>filter_clause</code></em>
evaluates to true are fed to the window function; other rows
are discarded. Only window functions that are aggregates accept
a <code class="literal">FILTER</code> clause.
</p><p>
The built-in window functions are described in <a class="xref" href="functions-window.html#FUNCTIONS-WINDOW-TABLE" title="Table 9.57. General-Purpose Window Functions">Table 9.57</a>. Other window functions can be added by
the user. Also, any built-in or user-defined general-purpose or
statistical aggregate can be used as a window function. (Ordered-set
and hypothetical-set aggregates cannot presently be used as window functions.)
</p><p>
The syntaxes using <code class="literal">*</code> are used for calling parameter-less
aggregate functions as window functions, for example
<code class="literal">count(*) OVER (PARTITION BY x ORDER BY y)</code>.
The asterisk (<code class="literal">*</code>) is customarily not used for
window-specific functions. Window-specific functions do not
allow <code class="literal">DISTINCT</code> or <code class="literal">ORDER BY</code> to be used within the
function argument list.
</p><p>
Window function calls are permitted only in the <code class="literal">SELECT</code>
list and the <code class="literal">ORDER BY</code> clause of the query.
</p><p>
More information about window functions can be found in
<a class="xref" href="tutorial-window.html" title="3.5. Window Functions">Section 3.5</a>,
<a class="xref" href="functions-window.html" title="9.21. Window Functions">Section 9.21</a>, and
<a class="xref" href="queries-table-expressions.html#QUERIES-WINDOW" title="7.2.5. Window Function Processing">Section 7.2.5</a>.
</p></div><div class="sect2" id="SQL-SYNTAX-TYPE-CASTS"><div class="titlepage"><div><div><h3 class="title">4.2.9. Type Casts</h3></div></div></div><a id="id-1.5.3.6.17.2" class="indexterm"></a><a id="id-1.5.3.6.17.3" class="indexterm"></a><a id="id-1.5.3.6.17.4" class="indexterm"></a><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 ( <em class="replaceable"><code>expression</code></em> AS <em class="replaceable"><code>type</code></em> )
<em class="replaceable"><code>expression</code></em>::<em class="replaceable"><code>type</code></em>
</pre><p>
The <code class="literal">CAST</code> syntax conforms to SQL; the syntax with
<code class="literal">::</code> 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 class="xref" href="sql-syntax-lexical.html#SQL-SYNTAX-CONSTANTS-GENERIC" title="4.1.2.7. Constants of Other Types">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">“<span class="quote">OK to apply implicitly</span>”</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">
<em class="replaceable"><code>typename</code></em> ( <em class="replaceable"><code>expression</code></em> )
</pre><p>
However, this only works for types whose names are also valid as
function names. For example, <code class="literal">double precision</code>
cannot be used this way, but the equivalent <code class="literal">float8</code>
can. Also, the names <code class="literal">interval</code>, <code class="literal">time</code>, and
<code class="literal">timestamp</code> 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"><h3 class="title">Note</h3><p>
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">“<span class="quote">function-like
syntax</span>”</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 class="xref" href="sql-createcast.html" title="CREATE CAST"><span class="refentrytitle">CREATE CAST</span></a>.
</p></div></div><div class="sect2" id="SQL-SYNTAX-COLLATE-EXPRS"><div class="titlepage"><div><div><h3 class="title">4.2.10. Collation Expressions</h3></div></div></div><a id="id-1.5.3.6.18.2" class="indexterm"></a><p>
The <code class="literal">COLLATE</code> clause overrides the collation of
an expression. It is appended to the expression it applies to:
</p><pre class="synopsis">
<em class="replaceable"><code>expr</code></em> COLLATE <em class="replaceable"><code>collation</code></em>
</pre><p>
where <em class="replaceable"><code>collation</code></em> is a possibly
schema-qualified identifier. The <code class="literal">COLLATE</code>
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 <code class="literal">COLLATE</code> clause are
overriding the sort order in an <code class="literal">ORDER BY</code> 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 <code class="literal">COLLATE</code> 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
<code class="literal">COLLATE</code> clause is attached to, because the collation that is
applied by the operator or function is derived by considering all
arguments, and an explicit <code class="literal">COLLATE</code> clause will override the
collations of all other arguments. (Attaching non-matching
<code class="literal">COLLATE</code> clauses to more than one argument, however, is an
error. For more details see <a class="xref" href="collation.html" title="23.2. Collation Support">Section 23.2</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
<code class="literal">></code> operator, which is of the non-collatable data type
<code class="type">boolean</code>.
</p></div><div class="sect2" id="SQL-SYNTAX-SCALAR-SUBQUERIES"><div class="titlepage"><div><div><h3 class="title">4.2.11. Scalar Subqueries</h3></div></div></div><a id="id-1.5.3.6.19.2" class="indexterm"></a><p>
A scalar subquery is an ordinary
<code class="command">SELECT</code> query in parentheses that returns exactly one
row with one column. (See <a class="xref" href="queries.html" title="Chapter 7. Queries">Chapter 7</a> for information about writing queries.)
The <code class="command">SELECT</code> 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 class="xref" href="functions-subquery.html" title="9.22. Subquery Expressions">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" id="SQL-SYNTAX-ARRAY-CONSTRUCTORS"><div class="titlepage"><div><div><h3 class="title">4.2.12. Array Constructors</h3></div></div></div><a id="id-1.5.3.6.20.2" class="indexterm"></a><a id="id-1.5.3.6.20.3" class="indexterm"></a><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 <code class="literal">ARRAY</code>, a left square bracket
<code class="literal">[</code>, a list of expressions (separated by commas) for the
array element values, and finally a right square bracket <code class="literal">]</code>.
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 <code class="literal">UNION</code> or
<code class="literal">CASE</code> constructs (see <a class="xref" href="typeconv-union-case.html" title="10.5. UNION, CASE, and Related Constructs">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 class="xref" href="sql-expressions.html#SQL-SYNTAX-TYPE-CASTS" title="4.2.9. 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 <code class="literal">ARRAY</code> 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 <code class="literal">ARRAY</code> 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-<code class="literal">ARRAY</code> 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 <code class="literal">ARRAY</code> 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)
SELECT ARRAY(SELECT ARRAY[i, i*2] FROM generate_series(1,5) AS a(i));
array
----------------------------------
{{1,2},{2,4},{3,6},{4,8},{5,10}}
(1 row)
</pre><p>
The subquery must return a single column.
If the subquery's output column is of a non-array type, 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.
If the subquery's output column is of an array type, the result will be
an array of the same type but one higher dimension; in this case all
the subquery rows must yield arrays of identical dimensionality, else
the result would not be rectangular.
</p><p>
The subscripts of an array value built with <code class="literal">ARRAY</code>
always begin with one. For more information about arrays, see
<a class="xref" href="arrays.html" title="8.15. Arrays">Section 8.15</a>.
</p></div><div class="sect2" id="SQL-SYNTAX-ROW-CONSTRUCTORS"><div class="titlepage"><div><div><h3 class="title">4.2.13. Row Constructors</h3></div></div></div><a id="id-1.5.3.6.21.2" class="indexterm"></a><a id="id-1.5.3.6.21.3" class="indexterm"></a><a id="id-1.5.3.6.21.4" class="indexterm"></a><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
<code class="literal">ROW</code>, 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 <code class="literal">ROW</code> is optional when there is more than one
expression in the list.
</p><p>
A row constructor can include the syntax
<em class="replaceable"><code>rowvalue</code></em><code class="literal">.*</code>,
which will be expanded to a list of the elements of the row value,
just as occurs when the <code class="literal">.*</code> syntax is used at the top level
of a <code class="command">SELECT</code> list (see <a class="xref" href="rowtypes.html#ROWTYPES-USAGE" title="8.16.5. Using Composite Types in Queries">Section 8.16.5</a>).
For example, if table <code class="literal">t</code> has
columns <code class="literal">f1</code> and <code class="literal">f2</code>, 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"><h3 class="title">Note</h3><p>
Before <span class="productname">PostgreSQL</span> 8.2, the
<code class="literal">.*</code> syntax was not expanded in row constructors, so
that writing <code class="literal">ROW(t.*, 42)</code> 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 <code class="literal">.*</code>, for instance
<code class="literal">ROW(t, 42)</code>.
</p></div><p>
By default, the value created by a <code class="literal">ROW</code> 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 <code class="command">CREATE TYPE AS</code>. 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
<code class="literal">IS NULL</code> or <code class="literal">IS NOT NULL</code>, 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 class="xref" href="functions-comparisons.html" title="9.23. Row and Array Comparisons">Section 9.23</a>.
Row constructors can also be used in connection with subqueries,
as discussed in <a class="xref" href="functions-subquery.html" title="9.22. Subquery Expressions">Section 9.22</a>.
</p></div><div class="sect2" id="SYNTAX-EXPRESS-EVAL"><div class="titlepage"><div><div><h3 class="title">4.2.14. Expression Evaluation Rules</h3></div></div></div><a id="id-1.5.3.6.22.2" class="indexterm"></a><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 <code class="literal">somefunc()</code> 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">“<span class="quote">short-circuiting</span>”</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 <code class="literal">WHERE</code> and <code class="literal">HAVING</code> clauses,
since those clauses are extensively reprocessed as part of
developing an execution plan. Boolean
expressions (<code class="literal">AND</code>/<code class="literal">OR</code>/<code class="literal">NOT</code> 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 <code class="literal">CASE</code>
construct (see <a class="xref" href="functions-conditional.html" title="9.17. Conditional Expressions">Section 9.17</a>) can be
used. For example, this is an untrustworthy way of trying to
avoid division by zero in a <code class="literal">WHERE</code> 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 <code class="literal">CASE</code> 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
<code class="literal">y > 1.5*x</code> instead.)
</p><p>
<code class="literal">CASE</code> 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 class="xref" href="xfunc-volatility.html" title="38.7. Function Volatility Categories">Section 38.7</a>, functions and
operators marked <code class="literal">IMMUTABLE</code> 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 <code class="literal">x > 0</code> so that the
<code class="literal">ELSE</code> 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
<code class="literal">IF</code>-<code class="literal">THEN</code>-<code class="literal">ELSE</code> statement to protect
a risky computation is much safer than just nesting it in a
<code class="literal">CASE</code> expression.
</p><p>
Another limitation of the same kind is that a <code class="literal">CASE</code> cannot
prevent evaluation of an aggregate expression contained within it,
because aggregate expressions are computed before other
expressions in a <code class="literal">SELECT</code> list or <code class="literal">HAVING</code> 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 <code class="structfield">employees</code> 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 <code class="literal">WHERE</code>
or <code class="literal">FILTER</code> clause to prevent problematic input rows from
reaching an aggregate function in the first place.
</p></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="sql-syntax-lexical.html">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="sql-syntax.html">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="sql-syntax-calling-funcs.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">4.1. Lexical Structure </td><td width="20%" align="center"><a accesskey="h" href="index.html">Home</a></td><td width="40%" align="right" valign="top"> 4.3. Calling Functions</td></tr></table></div></body></html>
