Transfer statistics during pg_upgrade.

Add support to pg_dump for dumping stats, and use that during
pg_upgrade so that statistics are transferred during upgrade. In most
cases this removes the need for a costly re-analyze after upgrade.

Some statistics are not transferred, such as extended statistics or
statistics with a custom stakind.

Now pg_dump accepts the options --schema-only, --no-schema,
--data-only, --no-data, --statistics-only, and --no-statistics; which
allow all combinations of schema, data, and/or stats. The options are
named this way to preserve compatibility with the previous
--schema-only and --data-only options.

Statistics are in SECTION_DATA, unless the object itself is in
SECTION_POST_DATA.

The stats are represented as calls to pg_restore_relation_stats() and
pg_restore_attribute_stats().

Author: Corey Huinker, Jeff Davis
Reviewed-by: Jian He
Discussion: https://postgr.es/m/CADkLM=fzX7QX6r78fShWDjNN3Vcr4PVAnvXxQ4DiGy6V=0bCUA@mail.gmail.com
Discussion: https://postgr.es/m/CADkLM%3DcB0rF3p_FuWRTMSV0983ihTRpsH%2BOCpNyiqE7Wk0vUWA%40mail.gmail.com

Author: jeff-davis

doc/src/sgml/ref/pg_dump.sgml

@@ -123,7 +123,7 @@ PostgreSQL documentation
       <term><option>--data-only</option></term>
       <listitem>
        <para>
-        Dump only the data, not the schema (data definitions).
+        Dump only the data, not the schema (data definitions) or statistics.
         Table data, large objects, and sequence values are dumped.
        </para>
 
@@ -141,13 +141,15 @@ PostgreSQL documentation
       <listitem>
        <para>
         Include large objects in the dump.  This is the default behavior
-        except when <option>--schema</option>, <option>--table</option>, or
-        <option>--schema-only</option> is specified.  The <option>-b</option>
-        switch is therefore only useful to add large objects to dumps
-        where a specific schema or table has been requested.  Note that
-        large objects are considered data and therefore will be included when
-        <option>--data-only</option> is used, but not
-        when <option>--schema-only</option> is.
+        except when <option>--schema</option>, <option>--table</option>,
+        <option>--schema-only</option>, <option>--statistics-only</option>, or
+        <option>--no-data</option> is specified.  The <option>-b</option>
+        switch is therefore only useful to add large objects to dumps where a
+        specific schema or table has been requested.  Note that large objects
+        are considered data and therefore will be included when
+        <option>--data-only</option> is used, but not when
+        <option>--schema-only</option> or <option>--statistics-only</option>
+        is.
        </para>
       </listitem>
      </varlistentry>
@@ -516,10 +518,11 @@ PostgreSQL documentation
       <term><option>--schema-only</option></term>
       <listitem>
        <para>
-        Dump only the object definitions (schema), not data.
+        Dump only the object definitions (schema), not data or statistics.
        </para>
        <para>
-        This option is the inverse of <option>--data-only</option>.
+        This option is mutually exclusive to <option>--data-only</option>
+        and <option>--statistics-only</option>.
         It is similar to, but for historical reasons not identical to,
         specifying
         <option>--section=pre-data --section=post-data</option>.
@@ -652,6 +655,17 @@ PostgreSQL documentation
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--statistics-only</option></term>
+      <listitem>
+       <para>
+        Dump only the statistics, not the schema (data definitions) or data.
+        Statistics for tables, materialized views, and indexes are dumped.
+       </para>
+
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>-Z <replaceable class="parameter">level</replaceable></option></term>
       <term><option>-Z <replaceable class="parameter">method</replaceable></option>[:<replaceable>detail</replaceable>]</term>
@@ -741,7 +755,8 @@ PostgreSQL documentation
       <term><option>--disable-triggers</option></term>
       <listitem>
        <para>
-        This option is relevant only when creating a data-only dump.
+        This option is relevant only when creating a dump that includes data
+        but does not include schema.
         It instructs <application>pg_dump</application> to include commands
         to temporarily disable triggers on the target tables while
         the data is restored.  Use this if you have referential
@@ -833,7 +848,8 @@ PostgreSQL documentation
         though you do not need the data in it.
        </para>
        <para>
-        To exclude data for all tables in the database, see <option>--schema-only</option>.
+        To exclude data for all tables in the database, see <option>--schema-only</option>
+        or <option>--statistics-only</option>.
        </para>
       </listitem>
      </varlistentry>
@@ -1080,6 +1096,15 @@ PostgreSQL documentation
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--no-data</option></term>
+      <listitem>
+       <para>
+        Do not dump data.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>--no-publications</option></term>
       <listitem>
@@ -1098,6 +1123,24 @@ PostgreSQL documentation
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--no-schema</option></term>
+      <listitem>
+       <para>
+        Do not dump schema (data definitions).
+       </para>
+      </listitem>
+     </varlistentry>
+
+     <varlistentry>
+      <term><option>--no-statistics</option></term>
+      <listitem>
+       <para>
+        Do not dump statistics.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>--no-subscriptions</option></term>
       <listitem>
@@ -1236,9 +1279,11 @@ PostgreSQL documentation
          </para>
          <para>
           The data section contains actual table data, large-object
-          contents, and sequence values.
+          contents, statitistics for tables and materialized views and
+          sequence values.
           Post-data items include definitions of indexes, triggers, rules,
-          and constraints other than validated check constraints.
+          statistics for indexes, and constraints other than validated check
+          constraints.
           Pre-data items include all other data definition items.
          </para>
        </listitem>
@@ -1581,7 +1626,7 @@ CREATE DATABASE foo WITH TEMPLATE template0;
   </para>
 
   <para>
-   When a data-only dump is chosen and the option <option>--disable-triggers</option>
+   When a dump without schema is chosen and the option <option>--disable-triggers</option>
    is used, <application>pg_dump</application> emits commands
    to disable triggers on user tables before inserting the data,
    and then commands to re-enable them after the data has been

doc/src/sgml/ref/pg_dumpall.sgml

@@ -81,7 +81,7 @@ PostgreSQL documentation
       <term><option>--data-only</option></term>
       <listitem>
        <para>
-        Dump only the data, not the schema (data definitions).
+        Dump only the data, not the schema (data definitions) or statistics.
        </para>
       </listitem>
      </varlistentry>
@@ -265,6 +265,16 @@ exclude database <replaceable class="parameter">PATTERN</replaceable>
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--statistics-only</option></term>
+      <listitem>
+       <para>
+        Dump only the statistics, not the schema (data definitions) or data.
+        Statistics for tables, materialized views, and indexes are dumped.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>--binary-upgrade</option></term>
       <listitem>
@@ -307,7 +317,7 @@ exclude database <replaceable class="parameter">PATTERN</replaceable>
       <term><option>--disable-triggers</option></term>
       <listitem>
        <para>
-        This option is relevant only when creating a data-only dump.
+        This option is relevant only when creating a dump with data and without schema.
         It instructs <application>pg_dumpall</application> to include commands
         to temporarily disable triggers on the target tables while
         the data is restored.  Use this if you have referential
@@ -422,6 +432,15 @@ exclude database <replaceable class="parameter">PATTERN</replaceable>
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--no-data</option></term>
+      <listitem>
+       <para>
+        Do not dump data.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>--no-publications</option></term>
       <listitem>
@@ -447,6 +466,15 @@ exclude database <replaceable class="parameter">PATTERN</replaceable>
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--no-schema</option></term>
+      <listitem>
+       <para>
+        Do not dump schema (data definitions).
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>--no-security-labels</option></term>
       <listitem>
@@ -456,6 +484,15 @@ exclude database <replaceable class="parameter">PATTERN</replaceable>
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--no-statistics</option></term>
+      <listitem>
+       <para>
+        Do not dump statistics.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>--no-subscriptions</option></term>
       <listitem>

doc/src/sgml/ref/pg_restore.sgml

@@ -94,7 +94,7 @@ PostgreSQL documentation
       <term><option>--data-only</option></term>
       <listitem>
        <para>
-        Restore only the data, not the schema (data definitions).
+        Restore only the data, not the schema (data definitions) or statistics.
         Table data, large objects, and sequence values are restored,
         if present in the archive.
        </para>
@@ -483,10 +483,11 @@ PostgreSQL documentation
         to the extent that schema entries are present in the archive.
        </para>
        <para>
-        This option is the inverse of <option>--data-only</option>.
+        This option is mutually exclusive of <option>--data-only</option>
+        and <option>--statistics-only</option>.
         It is similar to, but for historical reasons not identical to,
         specifying
-        <option>--section=pre-data --section=post-data</option>.
+        <option>--section=pre-data --section=post-data --no-statistics</option>.
        </para>
        <para>
         (Do not confuse this with the <option>--schema</option> option, which
@@ -599,6 +600,15 @@ PostgreSQL documentation
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--statistics-only</option></term>
+      <listitem>
+       <para>
+        Restore only the statistics, not schema (data definitions) or data.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>-1</option></term>
       <term><option>--single-transaction</option></term>
@@ -617,7 +627,7 @@ PostgreSQL documentation
       <term><option>--disable-triggers</option></term>
       <listitem>
        <para>
-        This option is relevant only when performing a data-only restore.
+        This option is relevant only when performing a restore without schema.
         It instructs <application>pg_restore</application> to execute commands
         to temporarily disable triggers on the target tables while
         the data is restored.  Use this if you have referential
@@ -681,6 +691,16 @@ PostgreSQL documentation
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--no-data</option></term>
+      <listitem>
+       <para>
+        Do not output commands to restore data, even if the archive
+        contains them.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>--no-data-for-failed-tables</option></term>
       <listitem>
@@ -713,6 +733,16 @@ PostgreSQL documentation
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--no-schema</option></term>
+      <listitem>
+       <para>
+        Do not output commands to restore schema (data definitions), even if
+        the archive contains them.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>--no-security-labels</option></term>
       <listitem>
@@ -723,6 +753,16 @@ PostgreSQL documentation
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--no-statistics</option></term>
+      <listitem>
+       <para>
+        Do not output commands to restore statistics, even if the archive
+        contains them.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>--no-subscriptions</option></term>
       <listitem>

doc/src/sgml/ref/pgupgrade.sgml

@@ -145,6 +145,15 @@ PostgreSQL documentation
       </listitem>
      </varlistentry>
 
+     <varlistentry>
+      <term><option>--no-statistics</option></term>
+      <listitem>
+       <para>
+        Do not restore statistics from the old cluster into the new cluster.
+       </para>
+      </listitem>
+     </varlistentry>
+
      <varlistentry>
       <term><option>-o</option> <replaceable class="parameter">options</replaceable></term>
       <term><option>--old-options</option> <replaceable class="parameter">options</replaceable></term>

src/bin/pg_dump/pg_backup.h

@@ -160,6 +160,7 @@ typedef struct _restoreOptions
 	/* flags derived from the user-settable flags */
 	bool		dumpSchema;
 	bool		dumpData;
+	bool		dumpStatistics;
 } RestoreOptions;
 
 typedef struct _dumpOptions
@@ -208,6 +209,7 @@ typedef struct _dumpOptions
 	/* flags derived from the user-settable flags */
 	bool		dumpSchema;
 	bool		dumpData;
+	bool		dumpStatistics;
 } DumpOptions;
 
 /*