Instead of having a configure-time DEFAULT_ATTSTATTARGET, store -1 in

attstattarget to indicate 'use the default'.  The default is now a GUC
variable default_statistics_target, and so may be changed on the fly.  Along
the way we gain the ability to have pg_dump dump the per-column statistics
target when it's not the default.  Patch by Neil Conway, with some kibitzing
from Tom Lane.

Author: tglsfdc

doc/src/sgml/catalogs.sgml

@@ -1,6 +1,6 @@
 <!--
  Documentation of the system catalogs, directed toward PostgreSQL developers
- $Header: /cvsroot/pgsql/doc/src/sgml/catalogs.sgml,v 2.49 2002/07/30 05:24:56 tgl Exp $
+ $Header: /cvsroot/pgsql/doc/src/sgml/catalogs.sgml,v 2.50 2002/07/31 17:19:49 tgl Exp $
  -->
 
 <chapter id="catalogs">
@@ -672,8 +672,9 @@
        of statistics accumulated for this column by
        <command>ANALYZE</command>.
        A zero value indicates that no statistics should be collected.
-       The exact meaning of positive values is data type-dependent.
-       For scalar data types, <structfield>attstattarget</structfield>
+       A negative value says to use the system default statistics target.
+       The exact meaning of positive values is datatype-dependent.
+       For scalar datatypes, <structfield>attstattarget</structfield>
        is both the target number of <quote>most common values</quote>
        to collect, and the target number of histogram bins to create.
       </entry>

doc/src/sgml/ref/alter_table.sgml

@@ -1,5 +1,5 @@
 <!--
-$Header: /cvsroot/pgsql/doc/src/sgml/ref/alter_table.sgml,v 1.46 2002/07/12 18:43:12 tgl Exp $
+$Header: /cvsroot/pgsql/doc/src/sgml/ref/alter_table.sgml,v 1.47 2002/07/31 17:19:50 tgl Exp $
 PostgreSQL documentation
 -->
 
@@ -218,6 +218,8 @@ ALTER TABLE <replaceable class="PARAMETER">table</replaceable>
       This form
       sets the per-column statistics-gathering target for subsequent
       <xref linkend="sql-analyze" endterm="sql-analyze-title"> operations.
+      The target can be set in the range 0 to 1000; alternatively, set it
+      to -1 to revert to using the system default statistics target.
      </para>
     </listitem>
    </varlistentry>

doc/src/sgml/ref/analyze.sgml

@@ -1,5 +1,5 @@
 <!--
-$Header: /cvsroot/pgsql/doc/src/sgml/ref/analyze.sgml,v 1.8 2002/04/23 02:07:15 tgl Exp $
+$Header: /cvsroot/pgsql/doc/src/sgml/ref/analyze.sgml,v 1.9 2002/07/31 17:19:51 tgl Exp $
 PostgreSQL documentation
 -->
 
@@ -157,16 +157,18 @@ ANALYZE [ VERBOSE ] [ <replaceable class="PARAMETER">table</replaceable> [ (<rep
   </para>
 
   <para>
-   The extent of analysis can be controlled by adjusting the per-column
+   The extent of analysis can be controlled by adjusting the
+   <literal>default_statistics_target</> parameter variable, or on a
+   column-by-column basis by setting the per-column
    statistics target with <command>ALTER TABLE ALTER COLUMN SET
    STATISTICS</command> (see
    <xref linkend="sql-altertable" endterm="sql-altertable-title">).  The
    target value sets the maximum number of entries in the most-common-value
    list and the maximum number of bins in the histogram.  The default
    target value is 10, but this can be adjusted up or down to trade off
    accuracy of planner estimates against the time taken for
-   <command>ANALYZE</command> and the
-   amount of space occupied in <literal>pg_statistic</literal>.
+   <command>ANALYZE</command> and the amount of space occupied
+   in <literal>pg_statistic</literal>.
    In particular, setting the statistics target to zero disables collection of
    statistics for that column.  It may be useful to do that for columns that
    are never used as part of the WHERE, GROUP BY, or ORDER BY clauses of

doc/src/sgml/runtime.sgml

@@ -1,5 +1,5 @@
 <!--
-$Header: /cvsroot/pgsql/doc/src/sgml/runtime.sgml,v 1.121 2002/07/13 01:02:14 momjian Exp $
+$Header: /cvsroot/pgsql/doc/src/sgml/runtime.sgml,v 1.122 2002/07/31 17:19:50 tgl Exp $
 -->
 
 <Chapter Id="runtime">
@@ -592,6 +592,19 @@ env PGOPTIONS='-c geqo=off' psql
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><varname>DEFAULT_STATISTICS_TARGET</varname> (<type>integer</type>)</term>
+      <listitem>
+       <para>
+        Sets the default statistics target for table columns that have not
+	had a column-specific target set via <command>ALTER TABLE SET
+	STATISTICS</>.  Larger values increase the time needed to do
+	<command>ANALYZE</>, but may improve the quality of the planner's
+	estimates.
+       </para>
+      </listitem>
+     </varlistentry>
+    
      <varlistentry>
       <term><varname>EFFECTIVE_CACHE_SIZE</varname> (<type>floating point</type>)</term>
       <listitem>

src/backend/access/common/tupdesc.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/access/common/tupdesc.c,v 1.81 2002/07/20 05:16:56 momjian Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/access/common/tupdesc.c,v 1.82 2002/07/31 17:19:51 tgl Exp $
  *
  * NOTES
  *	  some of the executor utility code such as "ExecTypeFromTL" should be
@@ -375,7 +375,7 @@ TupleDescInitEntry(TupleDesc desc,
 	else
 		MemSet(NameStr(att->attname), 0, NAMEDATALEN);
 
-	att->attstattarget = 0;
+	att->attstattarget = -1;
 	att->attcacheoff = -1;
 	att->atttypmod = typmod;
 

src/backend/catalog/genbki.sh

@@ -10,7 +10,7 @@
 #
 #
 # IDENTIFICATION
-#    $Header: /cvsroot/pgsql/src/backend/catalog/Attic/genbki.sh,v 1.27 2002/04/27 21:24:33 tgl Exp $
+#    $Header: /cvsroot/pgsql/src/backend/catalog/Attic/genbki.sh,v 1.28 2002/07/31 17:19:51 tgl Exp $
 #
 # NOTES
 #    non-essential whitespace is removed from the generated file.
@@ -126,12 +126,11 @@ for dir in $INCLUDE_DIRS; do
     fi
 done
 
-# Get INDEX_MAX_KEYS and DEFAULT_ATTSTATTARGET from pg_config.h
+# Get INDEX_MAX_KEYS from pg_config.h
 # (who needs consistency?)
 for dir in $INCLUDE_DIRS; do
     if [ -f "$dir/pg_config.h" ]; then
         INDEXMAXKEYS=`grep '^#define[ 	]*INDEX_MAX_KEYS' $dir/pg_config.h | $AWK '{ print $3 }'`
-        DEFAULTATTSTATTARGET=`grep '^#define[ 	]*DEFAULT_ATTSTATTARGET' $dir/pg_config.h | $AWK '{ print $3 }'`
         break
     fi
 done
@@ -194,7 +193,6 @@ sed -e "s/;[ 	]*$//g" \
     -e "s/PGUID/1/g" \
     -e "s/NAMEDATALEN/$NAMEDATALEN/g" \
     -e "s/PGNSP/$PG_CATALOG_NAMESPACE/g" \
-    -e "s/DEFAULT_ATTSTATTARGET/$DEFAULTATTSTATTARGET/g" \
     -e "s/INDEX_MAX_KEYS\*2/$INDEXMAXKEYS2/g" \
     -e "s/INDEX_MAX_KEYS\*4/$INDEXMAXKEYS4/g" \
     -e "s/INDEX_MAX_KEYS/$INDEXMAXKEYS/g" \

src/backend/catalog/heap.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/catalog/heap.c,v 1.213 2002/07/24 19:11:07 petere Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/catalog/heap.c,v 1.214 2002/07/31 17:19:51 tgl Exp $
  *
  *
  * INTERFACE ROUTINES
@@ -441,7 +441,7 @@ AddNewAttributeTuples(Oid new_rel_oid,
 		/* Fill in the correct relation OID */
 		(*dpp)->attrelid = new_rel_oid;
 		/* Make sure these are OK, too */
-		(*dpp)->attstattarget = DEFAULT_ATTSTATTARGET;
+		(*dpp)->attstattarget = -1;
 		(*dpp)->attcacheoff = -1;
 
 		tup = heap_addheader(Natts_pg_attribute,

src/backend/commands/analyze.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/commands/analyze.c,v 1.38 2002/06/20 20:29:26 momjian Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/commands/analyze.c,v 1.39 2002/07/31 17:19:51 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -107,6 +107,11 @@ typedef struct
 #define swapInt(a,b)	do {int _tmp; _tmp=a; a=b; b=_tmp;} while(0)
 #define swapDatum(a,b)	do {Datum _tmp; _tmp=a; a=b; b=_tmp;} while(0)
 
+
+/* Default statistics target (GUC parameter) */
+int		default_statistics_target = 10;
+
+
 static int elevel = -1;
 
 static MemoryContext anl_context = NULL;
@@ -384,7 +389,7 @@ examine_attribute(Relation onerel, int attnum)
 	VacAttrStats *stats;
 
 	/* Don't analyze column if user has specified not to */
-	if (attr->attstattarget <= 0)
+	if (attr->attstattarget == 0)
 		return NULL;
 
 	/* If column has no "=" operator, we can't do much of anything */
@@ -425,6 +430,10 @@ examine_attribute(Relation onerel, int attnum)
 	stats->eqopr = eqopr;
 	stats->eqfunc = eqfunc;
 
+	/* If the attstattarget column is negative, use the default value */
+	if (stats->attr->attstattarget < 0)
+		stats->attr->attstattarget = default_statistics_target;
+
 	/* Is there a "<" operator with suitable semantics? */
 	func_operator = compatible_oper(makeList1(makeString("<")),
 									attr->atttypid,
@@ -466,14 +475,14 @@ examine_attribute(Relation onerel, int attnum)
 		 * know it at this point.
 		 *--------------------
 		 */
-		stats->minrows = 300 * attr->attstattarget;
+		stats->minrows = 300 * stats->attr->attstattarget;
 	}
 	else
 	{
 		/* Can't do much but the minimal stuff */
 		stats->algcode = ALG_MINIMAL;
 		/* Might as well use the same minrows as above */
-		stats->minrows = 300 * attr->attstattarget;
+		stats->minrows = 300 * stats->attr->attstattarget;
 	}
 
 	return stats;

src/backend/commands/tablecmds.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/commands/tablecmds.c,v 1.24 2002/07/20 05:16:57 momjian Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/commands/tablecmds.c,v 1.25 2002/07/31 17:19:51 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -1668,7 +1668,7 @@ AlterTableAddColumn(Oid myrelid,
 	attribute->attrelid = myrelid;
 	namestrcpy(&(attribute->attname), colDef->colname);
 	attribute->atttypid = HeapTupleGetOid(typeTuple);
-	attribute->attstattarget = DEFAULT_ATTSTATTARGET;
+	attribute->attstattarget = -1;
 	attribute->attlen = tform->typlen;
 	attribute->attcacheoff = -1;
 	attribute->atttypmod = colDef->typename->typmod;
@@ -2184,12 +2184,18 @@ AlterTableAlterColumnFlags(Oid myrelid,
 		newtarget = intVal(flagValue);
 
 		/*
-		 * Limit target to sane range (should we raise an error instead?)
+		 * Limit target to a sane range
 		 */
-		if (newtarget < 0)
-			newtarget = 0;
+		if (newtarget < -1)
+		{
+			elog(ERROR, "ALTER TABLE: statistics target %d is too low",
+				 newtarget);
+		}
 		else if (newtarget > 1000)
+		{
+			elog(WARNING, "ALTER TABLE: lowering statistics target to 1000");
 			newtarget = 1000;
+		}
 	}
 	else if (*flagType == 'M')
 	{

src/backend/parser/gram.y

@@ -11,7 +11,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/parser/gram.y,v 2.351 2002/07/30 16:55:44 momjian Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/parser/gram.y,v 2.352 2002/07/31 17:19:51 tgl Exp $
  *
  * HISTORY
  *	  AUTHOR			DATE			MAJOR EVENT
@@ -1134,14 +1134,14 @@ AlterTableStmt:
 					n->name = $6;
 					$$ = (Node *)n;
 				}
-			/* ALTER TABLE <relation> ALTER [COLUMN] <colname> SET STATISTICS <Iconst> */
-			| ALTER TABLE relation_expr ALTER opt_column ColId SET STATISTICS Iconst
+			/* ALTER TABLE <relation> ALTER [COLUMN] <colname> SET STATISTICS <IntegerOnly> */
+			| ALTER TABLE relation_expr ALTER opt_column ColId SET STATISTICS IntegerOnly
 				{
 					AlterTableStmt *n = makeNode(AlterTableStmt);
 					n->subtype = 'S';
 					n->relation = $3;
 					n->name = $6;
-					n->def = (Node *) makeInteger($9);
+					n->def = (Node *) $9;
 					$$ = (Node *)n;
 				}
 			/* ALTER TABLE <relation> ALTER [COLUMN] <colname> SET STORAGE <storagemode> */

src/backend/utils/misc/guc.c

@@ -5,7 +5,7 @@
  * command, configuration file, and command line options.
  * See src/backend/utils/misc/README for more information.
  *
- * $Header: /cvsroot/pgsql/src/backend/utils/misc/guc.c,v 1.76 2002/07/30 16:20:03 momjian Exp $
+ * $Header: /cvsroot/pgsql/src/backend/utils/misc/guc.c,v 1.77 2002/07/31 17:19:52 tgl Exp $
  *
  * Copyright 2000 by PostgreSQL Global Development Group
  * Written by Peter Eisentraut <peter_e@gmx.net>.
@@ -26,6 +26,7 @@
 #include "catalog/pg_type.h"
 #include "commands/async.h"
 #include "commands/variable.h"
+#include "commands/vacuum.h"
 #include "executor/executor.h"
 #include "fmgr.h"
 #include "libpq/auth.h"
@@ -491,6 +492,10 @@ static struct config_bool
 static struct config_int
 			ConfigureNamesInt[] =
 {
+	{
+		{ "default_statistics_target", PGC_USERSET }, &default_statistics_target,
+		10, 1, 1000, NULL, NULL
+	},
 	{
 		{ "geqo_threshold", PGC_USERSET }, &geqo_rels,
 		DEFAULT_GEQO_RELS, 2, INT_MAX, NULL, NULL

src/backend/utils/misc/postgresql.conf.sample

@@ -95,6 +95,7 @@
 #cpu_index_tuple_cost = 0.001
 #cpu_operator_cost = 0.0025
 
+#default_statistics_target = 10	# range 1-1000
 
 #
 #	GEQO Optimizer Parameters