<para>
In strict SQL, <literal>GROUP BY</literal> can only group by columns of
- the source table but <productname>PostgreSQL</productname> extends
+ the source table, but <productname>PostgreSQL</productname> extends
this to also allow <literal>GROUP BY</literal> to group by columns in the
select list. Grouping by value expressions instead of simple
- column names is also allowed.
+ column names is also allowed (but <literal>GROUP BY</literal>
+ expressions cannot contain aggregate functions or window functions).
+ </para>
+
+ <para>
+ PostgreSQL also supports the syntax <literal>GROUP BY ALL</literal>,
+ which is equivalent to explicitly writing all select-list entries that
+ do not contain either an aggregate function or a window function.
+ This can greatly simplify ad-hoc exploration of data.
+ As an example, these queries are equivalent:
+<screen>
+<prompt>=></prompt> <userinput>SELECT a, b, a + b, sum(c) FROM test1 GROUP BY ALL;</userinput>
+ a | b | ?column? | sum
+---+---+----------+----
+ 1 | 4 | 5 | 9
+ 2 | 5 | 7 | 12
+ 3 | 6 | 9 | 15
+(3 rows)
+
+<prompt>=></prompt> <userinput>SELECT a, b, a + b, sum(c) FROM test1 GROUP BY a, b, a + b;</userinput>
+ a | b | ?column? | sum
+---+---+----------+----
+ 1 | 4 | 5 | 9
+ 2 | 5 | 7 | 12
+ 3 | 6 | 9 | 15
+(3 rows)
+</screen>
</para>
<indexterm>
[ { * | <replaceable class="parameter">expression</replaceable> [ [ AS ] <replaceable class="parameter">output_name</replaceable> ] } [, ...] ]
[ FROM <replaceable class="parameter">from_item</replaceable> [, ...] ]
[ WHERE <replaceable class="parameter">condition</replaceable> ]
- [ GROUP BY [ ALL | DISTINCT ] <replaceable class="parameter">grouping_element</replaceable> [, ...] ]
+ [ GROUP BY { ALL | [ ALL | DISTINCT ] <replaceable class="parameter">grouping_element</replaceable> [, ...] } ]
[ HAVING <replaceable class="parameter">condition</replaceable> ]
[ WINDOW <replaceable class="parameter">window_name</replaceable> AS ( <replaceable class="parameter">window_definition</replaceable> ) [, ...] ]
[ { UNION | INTERSECT | EXCEPT } [ ALL | DISTINCT ] <replaceable class="parameter">select</replaceable> ]
<para>
The optional <literal>GROUP BY</literal> clause has the general form
<synopsis>
-GROUP BY [ ALL | DISTINCT ] <replaceable class="parameter">grouping_element</replaceable> [, ...]
+GROUP BY { ALL | [ ALL | DISTINCT ] <replaceable class="parameter">grouping_element</replaceable> [, ...] }
</synopsis>
</para>
<replaceable class="parameter">grouping_element</replaceable>
can be an input column name, or the name or ordinal number of an
output column (<command>SELECT</command> list item), or an arbitrary
- expression formed from input-column values. In case of ambiguity,
+ expression formed from input-column values; however, it cannot contain
+ an aggregate function or a window function. In case of ambiguity,
a <literal>GROUP BY</literal> name will be interpreted as an
input-column name rather than an output column name.
</para>
+ <para>
+ The form <literal>GROUP BY ALL</literal> with no explicit
+ <replaceable class="parameter">grouping_elements</replaceable>
+ provided is equivalent to writing <literal>GROUP BY</literal> with the
+ numbers of all <command>SELECT</command> output columns that do not
+ contain either an aggregate function or a window function.
+ </para>
+
<para>
If any of <literal>GROUPING SETS</literal>, <literal>ROLLUP</literal> or
<literal>CUBE</literal> are present as grouping elements, then the
<literal>GROUP BY</literal> clause as a whole defines some number of
independent <replaceable>grouping sets</replaceable>. The effect of this is
equivalent to constructing a <literal>UNION ALL</literal> between
- subqueries with the individual grouping sets as their
+ subqueries having the individual grouping sets as their
<literal>GROUP BY</literal> clauses. The optional <literal>DISTINCT</literal>
- clause removes duplicate sets before processing; it does <emphasis>not</emphasis>
- transform the <literal>UNION ALL</literal> into a <literal>UNION DISTINCT</literal>.
+ key word removes duplicate grouping sets before processing; it does <emphasis>not</emphasis>
+ transform the implied <literal>UNION ALL</literal> into
+ a <literal>UNION DISTINCT</literal>.
For further details on the handling
of grouping sets see <xref linkend="queries-grouping-sets"/>.
</para>
INTO [ TEMPORARY | TEMP | UNLOGGED ] [ TABLE ] <replaceable class="parameter">new_table</replaceable>
[ FROM <replaceable class="parameter">from_item</replaceable> [, ...] ]
[ WHERE <replaceable class="parameter">condition</replaceable> ]
- [ GROUP BY <replaceable class="parameter">expression</replaceable> [, ...] ]
+ [ GROUP BY { ALL | [ ALL | DISTINCT ] <replaceable class="parameter">grouping_element</replaceable> [, ...] } ]
[ HAVING <replaceable class="parameter">condition</replaceable> ]
[ WINDOW <replaceable class="parameter">window_name</replaceable> AS ( <replaceable class="parameter">window_definition</replaceable> ) [, ...] ]
[ { UNION | INTERSECT | EXCEPT } [ ALL | DISTINCT ] <replaceable class="parameter">select</replaceable> ]
[ ORDER BY <replaceable class="parameter">expression</replaceable> [ ASC | DESC | USING <replaceable class="parameter">operator</replaceable> ] [ NULLS { FIRST | LAST } ] [, ...] ]
[ LIMIT { <replaceable class="parameter">count</replaceable> | ALL } ]
[ OFFSET <replaceable class="parameter">start</replaceable> [ ROW | ROWS ] ]
- [ FETCH { FIRST | NEXT } [ <replaceable class="parameter">count</replaceable> ] { ROW | ROWS } ONLY ]
- [ FOR { UPDATE | SHARE } [ OF <replaceable class="parameter">table_name</replaceable> [, ...] ] [ NOWAIT ] [...] ]
+ [ FETCH { FIRST | NEXT } [ <replaceable class="parameter">count</replaceable> ] { ROW | ROWS } { ONLY | WITH TIES } ]
+ [ FOR { UPDATE | NO KEY UPDATE | SHARE | KEY SHARE } [ OF <replaceable class="parameter">from_reference</replaceable> [, ...] ] [ NOWAIT | SKIP LOCKED ] [...] ]
</synopsis>
</refsynopsisdiv>
qry->groupClause = transformGroupClause(pstate,
stmt->groupClause,
+ stmt->groupByAll,
&qry->groupingSets,
&qry->targetList,
qry->sortClause,
EXPR_KIND_GROUP_BY,
false /* allow SQL92 rules */ );
qry->groupDistinct = stmt->groupDistinct;
+ qry->groupByAll = stmt->groupByAll;
if (stmt->distinctClause == NIL)
{
typedef struct GroupClause
{
bool distinct;
+ bool all;
List *list;
} GroupClause;
n->whereClause = $6;
n->groupClause = ($7)->list;
n->groupDistinct = ($7)->distinct;
+ n->groupByAll = ($7)->all;
n->havingClause = $8;
n->windowClause = $9;
$$ = (Node *) n;
n->whereClause = $6;
n->groupClause = ($7)->list;
n->groupDistinct = ($7)->distinct;
+ n->groupByAll = ($7)->all;
n->havingClause = $8;
n->windowClause = $9;
$$ = (Node *) n;
GroupClause *n = (GroupClause *) palloc(sizeof(GroupClause));
n->distinct = $3 == SET_QUANTIFIER_DISTINCT;
+ n->all = false;
n->list = $4;
$$ = n;
}
+ | GROUP_P BY ALL
+ {
+ GroupClause *n = (GroupClause *) palloc(sizeof(GroupClause));
+ n->distinct = false;
+ n->all = true;
+ n->list = NIL;
+ $$ = n;
+ }
| /*EMPTY*/
{
GroupClause *n = (GroupClause *) palloc(sizeof(GroupClause));
n->distinct = false;
+ n->all = false;
n->list = NIL;
$$ = n;
}
n->whereClause = $4;
n->groupClause = ($5)->list;
n->groupDistinct = ($5)->distinct;
+ n->groupByAll = ($5)->all;
n->havingClause = $6;
n->windowClause = $7;
n->sortClause = $8;
* GROUP BY items will be added to the targetlist (as resjunk columns)
* if not already present, so the targetlist must be passed by reference.
*
+ * If GROUP BY ALL is specified, the groupClause will be inferred to be all
+ * non-aggregate, non-window expressions in the targetlist.
+ *
* This is also used for window PARTITION BY clauses (which act almost the
* same, but are always interpreted per SQL99 rules).
*
*
* pstate ParseState
* grouplist clause to transform
+ * groupByAll is this a GROUP BY ALL statement?
* groupingSets reference to list to contain the grouping set tree
* targetlist reference to TargetEntry list
* sortClause ORDER BY clause (SortGroupClause nodes)
* useSQL99 SQL99 rather than SQL92 syntax
*/
List *
-transformGroupClause(ParseState *pstate, List *grouplist, List **groupingSets,
+transformGroupClause(ParseState *pstate, List *grouplist, bool groupByAll,
+ List **groupingSets,
List **targetlist, List *sortClause,
ParseExprKind exprKind, bool useSQL99)
{
bool hasGroupingSets = false;
Bitmapset *seen_local = NULL;
+ /* Handle GROUP BY ALL */
+ if (groupByAll)
+ {
+ /* There cannot have been any explicit grouplist items */
+ Assert(grouplist == NIL);
+
+ /* Iterate over targets, adding acceptable ones to the result list */
+ foreach_ptr(TargetEntry, tle, *targetlist)
+ {
+ /* Ignore junk TLEs */
+ if (tle->resjunk)
+ continue;
+
+ /*
+ * TLEs containing aggregates are not okay to add to GROUP BY
+ * (compare checkTargetlistEntrySQL92). But the SQL standard
+ * directs us to skip them, so it's fine.
+ */
+ if (pstate->p_hasAggs &&
+ contain_aggs_of_level((Node *) tle->expr, 0))
+ continue;
+
+ /*
+ * Likewise, TLEs containing window functions are not okay to add
+ * to GROUP BY. At this writing, the SQL standard is silent on
+ * what to do with them, but by analogy to aggregates we'll just
+ * skip them.
+ */
+ if (pstate->p_hasWindowFuncs &&
+ contain_windowfuncs((Node *) tle->expr))
+ continue;
+
+ /*
+ * Otherwise, add the TLE to the result using default sort/group
+ * semantics. We specify the parse location as the TLE's
+ * location, despite the comment for addTargetToGroupList
+ * discouraging that. The only other thing we could point to is
+ * the ALL keyword, which seems unhelpful when there are multiple
+ * TLEs.
+ */
+ result = addTargetToGroupList(pstate, tle,
+ result, *targetlist,
+ exprLocation((Node *) tle->expr));
+ }
+
+ /* If we found any acceptable targets, we're done */
+ if (result != NIL)
+ return result;
+
+ /*
+ * Otherwise, the SQL standard says to treat it like "GROUP BY ()".
+ * Build a representation of that, and let the rest of this function
+ * handle it.
+ */
+ grouplist = list_make1(makeGroupingSet(GROUPING_SET_EMPTY, NIL, -1));
+ }
+
/*
* Recursively flatten implicit RowExprs. (Technically this is only needed
* for GROUP BY, per the syntax rules for grouping sets, but we do it
true /* force SQL99 rules */ );
partitionClause = transformGroupClause(pstate,
windef->partitionClause,
+ false /* not GROUP BY ALL */ ,
NULL,
targetlist,
orderClause,
save_ingroupby = context->inGroupBy;
context->inGroupBy = true;
- if (query->groupingSets == NIL)
+ if (query->groupByAll)
+ appendStringInfoString(buf, "ALL");
+ else if (query->groupingSets == NIL)
{
sep = "";
foreach(l, query->groupClause)
*/
/* yyyymmddN */
-#define CATALOG_VERSION_NO 202509191
+#define CATALOG_VERSION_NO 202509291
#endif
List *returningList; /* return-values list (of TargetEntry) */
List *groupClause; /* a list of SortGroupClause's */
- bool groupDistinct; /* is the group by clause distinct? */
+ bool groupDistinct; /* was GROUP BY DISTINCT used? */
+ bool groupByAll; /* was GROUP BY ALL used? */
List *groupingSets; /* a list of GroupingSet's if present */
Node *whereClause; /* WHERE qualification */
List *groupClause; /* GROUP BY clauses */
bool groupDistinct; /* Is this GROUP BY DISTINCT? */
+ bool groupByAll; /* Is this GROUP BY ALL? */
Node *havingClause; /* HAVING conditional-expression */
List *windowClause; /* WINDOW window_name AS (...), ... */
ParseExprKind exprKind, const char *constructName,
LimitOption limitOption);
extern List *transformGroupClause(ParseState *pstate, List *grouplist,
+ bool groupByAll,
List **groupingSets,
List **targetlist, List *sortClause,
ParseExprKind exprKind, bool useSQL99);
drop table t3;
drop table p_t1;
--
+-- Test GROUP BY ALL
+--
+-- We don't care about the data here, just the proper transformation of the
+-- GROUP BY clause, so test some queries and verify the EXPLAIN plans.
+--
+CREATE TEMP TABLE t1 (
+ a int,
+ b int,
+ c int
+);
+-- basic example
+EXPLAIN (COSTS OFF) SELECT b, COUNT(*) FROM t1 GROUP BY ALL;
+ QUERY PLAN
+----------------------
+ HashAggregate
+ Group Key: b
+ -> Seq Scan on t1
+(3 rows)
+
+-- multiple columns, non-consecutive order
+EXPLAIN (COSTS OFF) SELECT a, SUM(b), b FROM t1 GROUP BY ALL;
+ QUERY PLAN
+----------------------
+ HashAggregate
+ Group Key: a, b
+ -> Seq Scan on t1
+(3 rows)
+
+-- multi columns, no aggregate
+EXPLAIN (COSTS OFF) SELECT a + b FROM t1 GROUP BY ALL;
+ QUERY PLAN
+----------------------
+ HashAggregate
+ Group Key: (a + b)
+ -> Seq Scan on t1
+(3 rows)
+
+-- check we detect a non-top-level aggregate
+EXPLAIN (COSTS OFF) SELECT a, SUM(b) + 4 FROM t1 GROUP BY ALL;
+ QUERY PLAN
+----------------------
+ HashAggregate
+ Group Key: a
+ -> Seq Scan on t1
+(3 rows)
+
+-- including grouped column is okay
+EXPLAIN (COSTS OFF) SELECT a, SUM(b) + a FROM t1 GROUP BY ALL;
+ QUERY PLAN
+----------------------
+ HashAggregate
+ Group Key: a
+ -> Seq Scan on t1
+(3 rows)
+
+-- including non-grouped column, not so much
+EXPLAIN (COSTS OFF) SELECT a, SUM(b) + c FROM t1 GROUP BY ALL;
+ERROR: column "t1.c" must appear in the GROUP BY clause or be used in an aggregate function
+LINE 1: EXPLAIN (COSTS OFF) SELECT a, SUM(b) + c FROM t1 GROUP BY AL...
+ ^
+-- all aggregates, should reduce to GROUP BY ()
+EXPLAIN (COSTS OFF) SELECT COUNT(a), SUM(b) FROM t1 GROUP BY ALL;
+ QUERY PLAN
+----------------------
+ Aggregate
+ Group Key: ()
+ -> Seq Scan on t1
+(3 rows)
+
+-- likewise with empty target list
+EXPLAIN (COSTS OFF) SELECT FROM t1 GROUP BY ALL;
+ QUERY PLAN
+-----------------------
+ Result
+ Replaces: Aggregate
+(2 rows)
+
+-- window functions are not to be included in GROUP BY, either
+EXPLAIN (COSTS OFF) SELECT a, COUNT(a) OVER (PARTITION BY a) FROM t1 GROUP BY ALL;
+ QUERY PLAN
+----------------------------------
+ WindowAgg
+ Window: w1 AS (PARTITION BY a)
+ -> Sort
+ Sort Key: a
+ -> HashAggregate
+ Group Key: a
+ -> Seq Scan on t1
+(7 rows)
+
+-- all cols
+EXPLAIN (COSTS OFF) SELECT *, count(*) FROM t1 GROUP BY ALL;
+ QUERY PLAN
+----------------------
+ HashAggregate
+ Group Key: a, b, c
+ -> Seq Scan on t1
+(3 rows)
+
+-- group by all with grouping element(s) (equivalent to GROUP BY's
+-- default behavior, explicit antithesis to GROUP BY DISTINCT)
+EXPLAIN (COSTS OFF) SELECT a, count(*) FROM t1 GROUP BY ALL a;
+ QUERY PLAN
+----------------------
+ HashAggregate
+ Group Key: a
+ -> Seq Scan on t1
+(3 rows)
+
+-- verify deparsing of GROUP BY ALL
+CREATE TEMP VIEW v1 AS SELECT b, COUNT(*) FROM t1 GROUP BY ALL;
+SELECT pg_get_viewdef('v1'::regclass);
+ pg_get_viewdef
+-----------------------
+ SELECT b, +
+ count(*) AS count+
+ FROM t1 +
+ GROUP BY ALL;
+(1 row)
+
+DROP VIEW v1;
+DROP TABLE t1;
+--
-- Test GROUP BY matching of join columns that are type-coerced due to USING
--
create temp table t1(f1 int, f2 int);
drop table t3;
drop table p_t1;
+--
+-- Test GROUP BY ALL
+--
+-- We don't care about the data here, just the proper transformation of the
+-- GROUP BY clause, so test some queries and verify the EXPLAIN plans.
+--
+
+CREATE TEMP TABLE t1 (
+ a int,
+ b int,
+ c int
+);
+
+-- basic example
+EXPLAIN (COSTS OFF) SELECT b, COUNT(*) FROM t1 GROUP BY ALL;
+
+-- multiple columns, non-consecutive order
+EXPLAIN (COSTS OFF) SELECT a, SUM(b), b FROM t1 GROUP BY ALL;
+
+-- multi columns, no aggregate
+EXPLAIN (COSTS OFF) SELECT a + b FROM t1 GROUP BY ALL;
+
+-- check we detect a non-top-level aggregate
+EXPLAIN (COSTS OFF) SELECT a, SUM(b) + 4 FROM t1 GROUP BY ALL;
+
+-- including grouped column is okay
+EXPLAIN (COSTS OFF) SELECT a, SUM(b) + a FROM t1 GROUP BY ALL;
+
+-- including non-grouped column, not so much
+EXPLAIN (COSTS OFF) SELECT a, SUM(b) + c FROM t1 GROUP BY ALL;
+
+-- all aggregates, should reduce to GROUP BY ()
+EXPLAIN (COSTS OFF) SELECT COUNT(a), SUM(b) FROM t1 GROUP BY ALL;
+
+-- likewise with empty target list
+EXPLAIN (COSTS OFF) SELECT FROM t1 GROUP BY ALL;
+
+-- window functions are not to be included in GROUP BY, either
+EXPLAIN (COSTS OFF) SELECT a, COUNT(a) OVER (PARTITION BY a) FROM t1 GROUP BY ALL;
+
+-- all cols
+EXPLAIN (COSTS OFF) SELECT *, count(*) FROM t1 GROUP BY ALL;
+
+-- group by all with grouping element(s) (equivalent to GROUP BY's
+-- default behavior, explicit antithesis to GROUP BY DISTINCT)
+EXPLAIN (COSTS OFF) SELECT a, count(*) FROM t1 GROUP BY ALL a;
+
+-- verify deparsing of GROUP BY ALL
+CREATE TEMP VIEW v1 AS SELECT b, COUNT(*) FROM t1 GROUP BY ALL;
+SELECT pg_get_viewdef('v1'::regclass);
+
+DROP VIEW v1;
+DROP TABLE t1;
+
--
-- Test GROUP BY matching of join columns that are type-coerced due to USING
--