Linux SRV-16 6.18.44-paas #1 SMP PREEMPT_DYNAMIC Thu Aug 13 10:08:12 UTC 2026 x86_64
/proc/212/root/opt/postgresql-9.6.10-201809281556/doc/html/
//proc/212/root/opt/postgresql-9.6.10-201809281556/doc/html/release-9-5.html
 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML
><HEAD
><TITLE
>Release 9.5</TITLE
><META
NAME="GENERATOR"
CONTENT="Modular DocBook HTML Stylesheet Version 1.79"><LINK
REV="MADE"
HREF="mailto:pgsql-docs@postgresql.org"><LINK
REL="HOME"
TITLE="PostgreSQL 9.6.10 Documentation"
HREF="index.html"><LINK
REL="UP"
TITLE="Release Notes"
HREF="release.html"><LINK
REL="PREVIOUS"
TITLE="Release 9.5.1"
HREF="release-9-5-1.html"><LINK
REL="NEXT"
TITLE="Release 9.4.19"
HREF="release-9-4-19.html"><LINK
REL="STYLESHEET"
TYPE="text/css"
HREF="stylesheet.css"><META
HTTP-EQUIV="Content-Type"
CONTENT="text/html; charset=ISO-8859-1"><META
NAME="creation"
CONTENT="2018-08-06T20:25:02"></HEAD
><BODY
CLASS="SECT1"
><DIV
CLASS="NAVHEADER"
><TABLE
SUMMARY="Header navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TH
COLSPAN="4"
ALIGN="center"
VALIGN="bottom"
><A
HREF="index.html"
>PostgreSQL 9.6.10 Documentation</A
></TH
></TR
><TR
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
TITLE="Release 9.5.1"
HREF="release-9-5-1.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
HREF="release.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="60%"
ALIGN="center"
VALIGN="bottom"
>Appendix E. Release Notes</TD
><TD
WIDTH="20%"
ALIGN="right"
VALIGN="top"
><A
TITLE="Release 9.4.19"
HREF="release-9-4-19.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><DIV
CLASS="SECT1"
><H1
CLASS="SECT1"
><A
NAME="RELEASE-9-5"
>E.26. Release 9.5</A
></H1
><DIV
CLASS="FORMALPARA"
><P
><B
>Release date: </B
>2016-01-07</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN136301"
>E.26.1. Overview</A
></H2
><P
>    Major enhancements in <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 9.5 include:
   </P
><P
></P
><UL
><LI
><P
>      Allow <A
HREF="sql-insert.html#SQL-ON-CONFLICT"
><TT
CLASS="COMMAND"
>INSERT</TT
></A
>s
      that would generate constraint conflicts to be turned into
      <TT
CLASS="COMMAND"
>UPDATE</TT
>s or ignored
     </P
></LI
><LI
><P
>      Add <TT
CLASS="LITERAL"
>GROUP BY</TT
> analysis features <A
HREF="queries-table-expressions.html#QUERIES-GROUPING-SETS"
><TT
CLASS="LITERAL"
>GROUPING SETS</TT
></A
>,
      <A
HREF="queries-table-expressions.html#QUERIES-GROUPING-SETS"
><TT
CLASS="LITERAL"
>CUBE</TT
></A
> and
      <A
HREF="queries-table-expressions.html#QUERIES-GROUPING-SETS"
><TT
CLASS="LITERAL"
>ROLLUP</TT
></A
>
     </P
></LI
><LI
><P
>      Add <A
HREF="ddl-rowsecurity.html"
>row-level security control</A
>
     </P
></LI
><LI
><P
>      Create mechanisms for tracking
      the <A
HREF="replication-origins.html"
>progress of replication</A
>,
      including methods for identifying the origin of individual changes
      during logical replication
     </P
></LI
><LI
><P
>      Add <A
HREF="brin.html"
>Block Range Indexes</A
> (<ACRONYM
CLASS="ACRONYM"
>BRIN</ACRONYM
>)
     </P
></LI
><LI
><P
>      Substantial performance improvements for sorting
     </P
></LI
><LI
><P
>      Substantial performance improvements for multi-CPU machines
     </P
></LI
></UL
><P
>    The above items are explained in more detail in the sections below.
   </P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN136335"
>E.26.2. Migration to Version 9.5</A
></H2
><P
>    A dump/restore using <A
HREF="app-pg-dumpall.html"
><SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
></A
>, or use
    of <A
HREF="pgupgrade.html"
><SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
></A
>, is required for those wishing to migrate
    data from any previous release.
   </P
><P
>    Version 9.5 contains a number of changes that may affect compatibility
    with previous releases.  Observe the following incompatibilities:
   </P
><P
></P
><UL
><LI
><P
>      Adjust <A
HREF="sql-syntax-lexical.html#SQL-PRECEDENCE"
>operator precedence</A
>
      to match the <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
> standard (Tom Lane)
     </P
><P
>      The precedence of <TT
CLASS="LITERAL"
>&lt;=</TT
>, <TT
CLASS="LITERAL"
>&gt;=</TT
>
      and <TT
CLASS="LITERAL"
>&lt;&gt;</TT
> has been reduced to match that of
      <TT
CLASS="LITERAL"
>&lt;</TT
>, <TT
CLASS="LITERAL"
>&gt;</TT
>
      and <TT
CLASS="LITERAL"
>=</TT
>.  The precedence of <TT
CLASS="LITERAL"
>IS</TT
> tests
      (e.g., <TT
CLASS="REPLACEABLE"
><I
>x</I
></TT
> <TT
CLASS="LITERAL"
>IS NULL</TT
>) has been reduced to be
      just below these six comparison operators.
      Also, multi-keyword operators beginning with <TT
CLASS="LITERAL"
>NOT</TT
> now have
      the precedence of their base operator (for example, <TT
CLASS="LITERAL"
>NOT
      BETWEEN</TT
> now has the same precedence as <TT
CLASS="LITERAL"
>BETWEEN</TT
>) whereas
      before they had inconsistent precedence, behaving like <TT
CLASS="LITERAL"
>NOT</TT
>
      with respect to their left operand but like their base operator with
      respect to their right operand.  The new configuration
      parameter <A
HREF="runtime-config-compatible.html#GUC-OPERATOR-PRECEDENCE-WARNING"
>operator_precedence_warning</A
> can be
      enabled to warn about queries in which these precedence changes result
      in different parsing choices.
     </P
></LI
><LI
><P
>      Change <A
HREF="app-pg-ctl.html"
><SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
></A
>'s default shutdown mode from
      <TT
CLASS="LITERAL"
>smart</TT
> to <TT
CLASS="LITERAL"
>fast</TT
> (Bruce Momjian)
     </P
><P
>      This means the default behavior will be to forcibly cancel existing
      database sessions, not simply wait for them to exit.
     </P
></LI
><LI
><P
>      Use assignment cast behavior for data type conversions
      in <SPAN
CLASS="APPLICATION"
>PL/pgSQL</SPAN
> assignments, rather than converting to and
      from text (Tom Lane)
     </P
><P
>      This change causes conversions of Booleans to strings to
      produce <TT
CLASS="LITERAL"
>true</TT
> or <TT
CLASS="LITERAL"
>false</TT
>, not <TT
CLASS="LITERAL"
>t</TT
>
      or <TT
CLASS="LITERAL"
>f</TT
>.  Other type conversions may succeed in more cases
      than before; for example, assigning a numeric value <TT
CLASS="LITERAL"
>3.9</TT
> to
      an integer variable will now assign 4 rather than failing.  If no
      assignment-grade cast is defined for the particular source and
      destination types, <SPAN
CLASS="APPLICATION"
>PL/pgSQL</SPAN
> will fall back to its old
      I/O conversion behavior.
     </P
></LI
><LI
><P
>      Allow characters in <A
HREF="libpq-connect.html#LIBPQ-CONNECT-OPTIONS"
>server
      command-line options</A
> to be escaped with a backslash (Andres Freund)
     </P
><P
>      Formerly, spaces in the options string always separated options, so
      there was no way to include a space in an option value.  Including
      a backslash in an option value now requires writing <TT
CLASS="LITERAL"
>\\</TT
>.
     </P
></LI
><LI
><P
>      Change the default value of the GSSAPI <A
HREF="auth-methods.html#GSSAPI-AUTH"
><TT
CLASS="VARNAME"
>include_realm</TT
></A
> parameter to 1, so
      that by default the realm is not removed from a <ACRONYM
CLASS="ACRONYM"
>GSS</ACRONYM
>
      or <ACRONYM
CLASS="ACRONYM"
>SSPI</ACRONYM
> principal name (Stephen Frost)
     </P
></LI
><LI
><P
>      Replace configuration parameter <TT
CLASS="VARNAME"
>checkpoint_segments</TT
>
      with <A
HREF="runtime-config-wal.html#GUC-MIN-WAL-SIZE"
>min_wal_size</A
>
      and <A
HREF="runtime-config-wal.html#GUC-MAX-WAL-SIZE"
>max_wal_size</A
> (Heikki Linnakangas)
     </P
><P
>      If you previously adjusted <TT
CLASS="VARNAME"
>checkpoint_segments</TT
>, the
      following formula will give you an approximately equivalent setting:
</P><PRE
CLASS="PROGRAMLISTING"
>max_wal_size = (3 * checkpoint_segments) * 16MB</PRE
><P>
      Note that the default setting for <TT
CLASS="VARNAME"
>max_wal_size</TT
> is
      much higher than the default <TT
CLASS="VARNAME"
>checkpoint_segments</TT
> used
      to be, so adjusting it might no longer be necessary.
     </P
></LI
><LI
><P
>        Control the Linux <ACRONYM
CLASS="ACRONYM"
>OOM</ACRONYM
> killer via new environment
        variables <A
HREF="kernel-resources.html#LINUX-MEMORY-OVERCOMMIT"
><TT
CLASS="ENVAR"
>PG_OOM_ADJUST_FILE</TT
></A
>
        and <A
HREF="kernel-resources.html#LINUX-MEMORY-OVERCOMMIT"
><TT
CLASS="ENVAR"
>PG_OOM_ADJUST_VALUE</TT
></A
>,
        instead of compile-time options <TT
CLASS="LITERAL"
>LINUX_OOM_SCORE_ADJ</TT
> and
        <TT
CLASS="LITERAL"
>LINUX_OOM_ADJ</TT
>
        (Gurjeet Singh)
       </P
></LI
><LI
><P
>      Decommission server configuration
      parameter <TT
CLASS="VARNAME"
>ssl_renegotiation_limit</TT
>, which was deprecated
      in earlier releases (Andres Freund)
     </P
><P
>      While SSL renegotiation is a good idea in theory, it has caused enough
      bugs to be considered a net negative in practice, and it is due to be
      removed from future versions of the relevant standards.  We have
      therefore removed support for it from <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>.
      The <TT
CLASS="VARNAME"
>ssl_renegotiation_limit</TT
> parameter still exists, but
      cannot be set to anything but zero (disabled).  It's not documented
      anymore, either.
     </P
></LI
><LI
><P
>      Remove server configuration parameter <TT
CLASS="VARNAME"
>autocommit</TT
>, which
      was already deprecated and non-operational (Tom Lane)
     </P
></LI
><LI
><P
>      Remove the <A
HREF="catalog-pg-authid.html"
><TT
CLASS="STRUCTNAME"
>pg_authid</TT
></A
>
      catalog's <TT
CLASS="STRUCTFIELD"
>rolcatupdate</TT
> field, as it had no usefulness
      (Adam Brightwell)
     </P
></LI
><LI
><P
>      The <A
HREF="monitoring-stats.html#PG-STAT-REPLICATION-VIEW"
><TT
CLASS="STRUCTNAME"
>pg_stat_replication</TT
></A
>
      system view's <TT
CLASS="STRUCTFIELD"
>sent</TT
> field is now NULL, not zero, when
      it has no valid value (Magnus Hagander)
     </P
></LI
><LI
><P
>      Allow <TT
CLASS="TYPE"
>json</TT
> and <TT
CLASS="TYPE"
>jsonb</TT
> array extraction operators to
      accept negative subscripts, which count from the end of JSON arrays
      (Peter Geoghegan, Andrew Dunstan)
     </P
><P
>      Previously, these operators returned <TT
CLASS="LITERAL"
>NULL</TT
> for negative
      subscripts.
     </P
></LI
></UL
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN136432"
>E.26.3. Changes</A
></H2
><P
>    Below you will find a detailed account of the changes between
    <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 9.5 and the previous major
    release.
   </P
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN136436"
>E.26.3.1. Server</A
></H3
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN136438"
>E.26.3.1.1. Indexes</A
></H4
><P
></P
><UL
><LI
><P
>        Add <A
HREF="brin.html"
>Block Range Indexes</A
> (<ACRONYM
CLASS="ACRONYM"
>BRIN</ACRONYM
>)
        (&Aacute;lvaro Herrera)
       </P
><P
>        <ACRONYM
CLASS="ACRONYM"
>BRIN</ACRONYM
> indexes store only summary data (such as minimum
        and maximum values) for ranges of heap blocks.  They are therefore
        very compact and cheap to update; but if the data is naturally
        clustered, they can still provide substantial speedup of searches.
       </P
></LI
><LI
><P
>        Allow queries to perform accurate distance filtering of
        bounding-box-indexed objects (polygons, circles) using <A
HREF="gist.html"
>GiST</A
> indexes (Alexander Korotkov, Heikki
        Linnakangas)
       </P
><P
>        Previously, to exploit such an index a subquery had to be used to
        select a large number of rows ordered by bounding-box distance, and
        the result then had to be filtered further with a more accurate
        distance calculation.
       </P
></LI
><LI
><P
>        Allow <A
HREF="gist.html"
>GiST</A
> indexes to perform index-only
        scans (Anastasia Lubennikova, Heikki Linnakangas, Andreas Karlsson)
       </P
></LI
><LI
><P
>        Add configuration parameter <A
HREF="runtime-config-client.html#GUC-GIN-PENDING-LIST-LIMIT"
>gin_pending_list_limit</A
>
        to control the size of <ACRONYM
CLASS="ACRONYM"
>GIN</ACRONYM
> pending lists (Fujii Masao)
       </P
><P
>        This value can also be set on a per-index basis as an index storage
        parameter.  Previously the pending-list size was controlled
        by <A
HREF="runtime-config-resource.html#GUC-WORK-MEM"
>work_mem</A
>, which was awkward because
        appropriate values for <TT
CLASS="VARNAME"
>work_mem</TT
> are often much too large
        for this purpose.
       </P
></LI
><LI
><P
>        Issue a warning during the creation of <A
HREF="indexes-types.html"
>hash</A
> indexes because they are not
        crash-safe (Bruce Momjian)
       </P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN136464"
>E.26.3.1.2. General Performance</A
></H4
><P
></P
><UL
><LI
><P
>        Improve the speed of sorting of <TT
CLASS="TYPE"
>varchar</TT
>, <TT
CLASS="TYPE"
>text</TT
>,
        and <TT
CLASS="TYPE"
>numeric</TT
> fields via <SPAN
CLASS="QUOTE"
>"abbreviated"</SPAN
> keys
        (Peter Geoghegan, Andrew Gierth, Robert Haas)
       </P
></LI
><LI
><P
>        Extend the infrastructure that allows sorting to be performed by
        inlined, non-<ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
>-callable comparison functions to
        cover <TT
CLASS="COMMAND"
>CREATE INDEX</TT
>, <TT
CLASS="COMMAND"
>REINDEX</TT
>, and
        <TT
CLASS="COMMAND"
>CLUSTER</TT
> (Peter Geoghegan)
       </P
></LI
><LI
><P
>        Improve performance of hash joins (Tomas Vondra, Robert Haas)
       </P
></LI
><LI
><P
>        Improve concurrency of shared buffer replacement
        (Robert Haas, Amit Kapila, Andres Freund)
       </P
></LI
><LI
><P
>        Reduce the number of page locks and pins during index scans (Kevin Grittner)
       </P
><P
>        The primary benefit of this is to allow index vacuums to be blocked
        less often.
       </P
></LI
><LI
><P
>        Make per-backend tracking of buffer pins more memory-efficient
        (Andres Freund)
       </P
></LI
><LI
><P
>        Improve lock scalability (Andres Freund)
       </P
><P
>        This particularly addresses scalability problems when running on
        systems with multiple <ACRONYM
CLASS="ACRONYM"
>CPU</ACRONYM
> sockets.
       </P
></LI
><LI
><P
>        Allow the optimizer to remove unnecessary references to left-joined
        subqueries (David Rowley)
       </P
></LI
><LI
><P
>        Allow pushdown of query restrictions into subqueries with <A
HREF="tutorial-window.html"
>window functions</A
>, where appropriate
        (David Rowley)
       </P
></LI
><LI
><P
>        Allow a non-leakproof function to be pushed down into a security
        barrier view if the function does not receive any view output
        columns (Dean Rasheed)
       </P
></LI
><LI
><P
>        Teach the planner to use statistics obtained from an expression
        index on a boolean-returning function, when a matching function call
        appears in <TT
CLASS="LITERAL"
>WHERE</TT
> (Tom Lane)
       </P
></LI
><LI
><P
>        Make <TT
CLASS="COMMAND"
>ANALYZE</TT
> compute basic statistics (null fraction and
        average column width) even for columns whose data type lacks an
        equality function (Oleksandr Shulgin)
       </P
></LI
><LI
><P
>        Speed up <ACRONYM
CLASS="ACRONYM"
>CRC</ACRONYM
> (cyclic redundancy check) computations
        and switch to CRC-32C (Abhijit Menon-Sen, Heikki Linnakangas)
       </P
></LI
><LI
><P
>        Improve bitmap index scan performance (Teodor Sigaev, Tom Lane)
       </P
></LI
><LI
><P
>        Speed up <TT
CLASS="COMMAND"
>CREATE INDEX</TT
> by avoiding unnecessary memory
        copies (Robert Haas)
       </P
></LI
><LI
><P
>        Increase the number of buffer mapping partitions (Amit Kapila,
        Andres Freund, Robert Haas)
       </P
><P
>        This improves performance for highly concurrent workloads.
       </P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN136516"
>E.26.3.1.3. Monitoring</A
></H4
><P
></P
><UL
><LI
><P
>        Add per-table autovacuum logging control via new
        <TT
CLASS="VARNAME"
>log_autovacuum_min_duration</TT
> storage parameter
        (Michael Paquier)
       </P
></LI
><LI
><P
>        Add new configuration parameter <A
HREF="runtime-config-logging.html#GUC-CLUSTER-NAME"
>cluster_name</A
>
        (Thomas Munro)
       </P
><P
>        This string, typically set in <A
HREF="config-setting.html#CONFIG-SETTING-CONFIGURATION-FILE"
><TT
CLASS="FILENAME"
>postgresql.conf</TT
></A
>,
        allows clients to identify the cluster.  This name also appears
        in the process title of all server processes, allowing for easier
        identification of processes belonging to the same cluster.
       </P
></LI
><LI
><P
>        Prevent non-superusers from changing <A
HREF="runtime-config-logging.html#GUC-LOG-DISCONNECTIONS"
>log_disconnections</A
> on connection startup (Fujii Masao)
       </P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN136531"
>E.26.3.1.4. <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
></A
></H4
><P
></P
><UL
><LI
><P
>        Check <A
HREF="libpq-ssl.html"
><SPAN
CLASS="QUOTE"
>"Subject Alternative
        Names"</SPAN
></A
> in <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> server certificates, if present
        (Alexey Klyukin)
       </P
><P
>        When they are present, this replaces checks against the certificate's
        <SPAN
CLASS="QUOTE"
>"Common Name"</SPAN
>.
       </P
></LI
><LI
><P
>        Add system view <A
HREF="monitoring-stats.html#PG-STAT-SSL-VIEW"
><TT
CLASS="STRUCTNAME"
>pg_stat_ssl</TT
></A
> to report
        <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> connection information (Magnus Hagander)
       </P
></LI
><LI
><P
>        Add <SPAN
CLASS="APPLICATION"
>libpq</SPAN
> functions to return <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
>
        information in an implementation-independent way (Heikki Linnakangas)
       </P
><P
>        While <A
HREF="libpq-status.html#LIBPQ-PQGETSSL"
><CODE
CLASS="FUNCTION"
>PQgetssl()</CODE
></A
> can
        still be used to call <SPAN
CLASS="PRODUCTNAME"
>OpenSSL</SPAN
> functions, it is now
        considered deprecated because future versions
        of <SPAN
CLASS="APPLICATION"
>libpq</SPAN
> might support other <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
>
        implementations.  When possible, use the new
        functions <A
HREF="libpq-status.html#LIBPQ-PQSSLATTRIBUTE"
><CODE
CLASS="FUNCTION"
>PQsslAttribute()</CODE
></A
>, <A
HREF="libpq-status.html#LIBPQ-PQSSLATTRIBUTENAMES"
><CODE
CLASS="FUNCTION"
>PQsslAttributeNames()</CODE
></A
>,
        and <A
HREF="libpq-status.html#LIBPQ-PQSSLINUSE"
><CODE
CLASS="FUNCTION"
>PQsslInUse()</CODE
></A
>
        to obtain <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> information in
        an <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
>-implementation-independent way.
       </P
></LI
><LI
><P
>        Make <SPAN
CLASS="APPLICATION"
>libpq</SPAN
> honor any <SPAN
CLASS="PRODUCTNAME"
>OpenSSL</SPAN
>
        thread callbacks (Jan Urbanski)
       </P
><P
>        Previously they were overwritten.
       </P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN136570"
>E.26.3.1.5. Server Settings</A
></H4
><P
></P
><UL
><LI
><P
>        Replace configuration parameter <TT
CLASS="VARNAME"
>checkpoint_segments</TT
>
        with <A
HREF="runtime-config-wal.html#GUC-MIN-WAL-SIZE"
>min_wal_size</A
>
        and <A
HREF="runtime-config-wal.html#GUC-MAX-WAL-SIZE"
>max_wal_size</A
> (Heikki Linnakangas)
       </P
><P
>        This change allows the allocation of a large number of <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
>
        files without keeping them after they are no longer needed.
        Therefore the default for <TT
CLASS="VARNAME"
>max_wal_size</TT
> has been set
        to <TT
CLASS="LITERAL"
>1GB</TT
>, much larger than the old default
        for <TT
CLASS="VARNAME"
>checkpoint_segments</TT
>.
        Also note that standby servers perform restartpoints to try to limit
        their WAL space consumption to <TT
CLASS="VARNAME"
>max_wal_size</TT
>; previously
        they did not pay any attention to <TT
CLASS="VARNAME"
>checkpoint_segments</TT
>.
       </P
></LI
><LI
><P
>        Control the Linux <ACRONYM
CLASS="ACRONYM"
>OOM</ACRONYM
> killer via new environment
        variables <A
HREF="kernel-resources.html#LINUX-MEMORY-OVERCOMMIT"
><TT
CLASS="ENVAR"
>PG_OOM_ADJUST_FILE</TT
></A
>
        and <A
HREF="kernel-resources.html#LINUX-MEMORY-OVERCOMMIT"
><TT
CLASS="ENVAR"
>PG_OOM_ADJUST_VALUE</TT
></A
>
        (Gurjeet Singh)
       </P
><P
>        The previous <ACRONYM
CLASS="ACRONYM"
>OOM</ACRONYM
> control infrastructure involved
        compile-time options <TT
CLASS="LITERAL"
>LINUX_OOM_SCORE_ADJ</TT
> and
        <TT
CLASS="LITERAL"
>LINUX_OOM_ADJ</TT
>, which are no longer supported.
        The new behavior is available in all builds.
       </P
></LI
><LI
><P
>        Allow recording of transaction
        commit time stamps when configuration parameter <A
HREF="runtime-config-replication.html#GUC-TRACK-COMMIT-TIMESTAMP"
>track_commit_timestamp</A
>
        is enabled (&Aacute;lvaro Herrera, Petr Jel&iacute;nek)
       </P
><P
>        Time stamp information can be accessed using functions <A
HREF="functions-info.html#FUNCTIONS-COMMIT-TIMESTAMP"
><CODE
CLASS="FUNCTION"
>pg_xact_commit_timestamp()</CODE
></A
>
        and <CODE
CLASS="FUNCTION"
>pg_last_committed_xact()</CODE
>.
       </P
></LI
><LI
><P
>        Allow <A
HREF="runtime-config-client.html#GUC-LOCAL-PRELOAD-LIBRARIES"
>local_preload_libraries</A
> to be set
        by <TT
CLASS="COMMAND"
>ALTER ROLE SET</TT
> (Peter Eisentraut, Kyotaro Horiguchi)
       </P
></LI
><LI
><P
>        Allow <A
HREF="routine-vacuuming.html#AUTOVACUUM"
>autovacuum workers</A
>
        to respond to configuration parameter changes during a run
        (Michael Paquier)
       </P
></LI
><LI
><P
>        Make configuration parameter <A
HREF="runtime-config-preset.html#GUC-DEBUG-ASSERTIONS"
>debug_assertions</A
>
        read-only (Andres Freund)
       </P
><P
>        This means that assertions can no longer be turned
        off if they were enabled at compile time, allowing for more
        efficient code optimization.  This change also removes the <A
HREF="app-postgres.html#APP-POSTGRES-OPTIONS"
>postgres</A
> <TT
CLASS="OPTION"
>-A</TT
> option.
       </P
></LI
><LI
><P
>        Allow setting <A
HREF="runtime-config-resource.html#GUC-EFFECTIVE-IO-CONCURRENCY"
>effective_io_concurrency</A
> on
        systems where it has no effect (Peter Eisentraut)
       </P
></LI
><LI
><P
>        Add system view <A
HREF="view-pg-file-settings.html"
><TT
CLASS="STRUCTNAME"
>pg_file_settings</TT
></A
>
        to show the contents of the server's configuration files
        (Sawada Masahiko)
       </P
></LI
><LI
><P
>        Add <TT
CLASS="STRUCTNAME"
>pending_restart</TT
> to the system view <A
HREF="view-pg-settings.html"
><TT
CLASS="STRUCTNAME"
>pg_settings</TT
></A
> to
        indicate a change has been made but will not take effect until a
        database restart (Peter Eisentraut)
       </P
></LI
><LI
><P
>        Allow <A
HREF="sql-altersystem.html"
><TT
CLASS="COMMAND"
>ALTER SYSTEM</TT
></A
>
        values to be reset with <TT
CLASS="COMMAND"
>ALTER SYSTEM RESET</TT
> (Vik
        Fearing)
       </P
><P
>        This command removes the specified setting
        from <TT
CLASS="FILENAME"
>postgresql.auto.conf</TT
>.
       </P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN136635"
>E.26.3.2. Replication and Recovery</A
></H3
><P
></P
><UL
><LI
><P
>        Create mechanisms for tracking
        the <A
HREF="replication-origins.html"
>progress of replication</A
>,
        including methods for identifying the origin of individual changes
        during logical replication (Andres Freund)
       </P
><P
>        This is helpful when implementing replication solutions.
       </P
></LI
><LI
><P
>        Rework truncation of the multixact commit log to be properly
        WAL-logged (Andres Freund)
       </P
><P
>        This makes things substantially simpler and more robust.
       </P
></LI
><LI
><P
>        Add <A
HREF="recovery-config.html"
><TT
CLASS="FILENAME"
>recovery.conf</TT
></A
>
        parameter <A
HREF="recovery-target-settings.html#RECOVERY-TARGET-ACTION"
><TT
CLASS="VARNAME"
>recovery_target_action</TT
></A
>
        to control post-recovery activity (Petr Jel&iacute;nek)
       </P
><P
>        This replaces the old parameter <TT
CLASS="VARNAME"
>pause_at_recovery_target</TT
>.
       </P
></LI
><LI
><P
>        Add new <A
HREF="runtime-config-wal.html#GUC-ARCHIVE-MODE"
>archive_mode</A
> value
        <TT
CLASS="LITERAL"
>always</TT
> to allow standbys to always archive received
        <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> files (Fujii Masao)
       </P
></LI
><LI
><P
>        Add configuration
        parameter <A
HREF="runtime-config-replication.html#GUC-WAL-RETRIEVE-RETRY-INTERVAL"
>wal_retrieve_retry_interval</A
> to
        control <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> read retry after failure
        (Alexey Vasiliev, Michael Paquier)
       </P
><P
>        This is particularly helpful for warm standbys.
       </P
></LI
><LI
><P
>        Allow compression of full-page images stored in <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
>
        (Rahila Syed, Michael Paquier)
       </P
><P
>        This feature reduces WAL volume, at the cost of more CPU time spent
        on WAL logging and WAL replay.  It is controlled by a new
        configuration parameter <A
HREF="runtime-config-wal.html#GUC-WAL-COMPRESSION"
>wal_compression</A
>, which
        currently is off by default.
       </P
></LI
><LI
><P
>        Archive <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> files with suffix <TT
CLASS="LITERAL"
>.partial</TT
>
        during standby promotion (Heikki Linnakangas)
       </P
></LI
><LI
><P
>        Add configuration parameter <A
HREF="runtime-config-logging.html#GUC-LOG-REPLICATION-COMMANDS"
>log_replication_commands</A
>
        to log replication commands (Fujii Masao)
       </P
><P
>        By default, replication commands, e.g. <A
HREF="protocol-replication.html"
><TT
CLASS="LITERAL"
>IDENTIFY_SYSTEM</TT
></A
>,
        are not logged, even when <A
HREF="runtime-config-logging.html#GUC-LOG-STATEMENT"
>log_statement</A
> is set
        to <TT
CLASS="LITERAL"
>all</TT
>.
       </P
></LI
><LI
><P
>        Report the processes holding replication slots in <A
HREF="view-pg-replication-slots.html"
><TT
CLASS="STRUCTNAME"
>pg_replication_slots</TT
></A
>
        (Craig Ringer)
       </P
><P
>        The new output column is <TT
CLASS="STRUCTNAME"
>active_pid</TT
>.
       </P
></LI
><LI
><P
>        Allow <TT
CLASS="FILENAME"
>recovery.conf</TT
>'s <A
HREF="standby-settings.html#PRIMARY-CONNINFO"
><TT
CLASS="VARNAME"
>primary_conninfo</TT
></A
> setting to
        use connection <ACRONYM
CLASS="ACRONYM"
>URI</ACRONYM
>s, e.g. <TT
CLASS="LITERAL"
>postgres://</TT
>
        (Alexander Shulgin)
       </P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN136693"
>E.26.3.3. Queries</A
></H3
><P
></P
><UL
><LI
><P
>        Allow <A
HREF="sql-insert.html#SQL-ON-CONFLICT"
><TT
CLASS="COMMAND"
>INSERT</TT
></A
>s
        that would generate constraint conflicts to be turned into
        <TT
CLASS="COMMAND"
>UPDATE</TT
>s or ignored (Peter Geoghegan, Heikki
        Linnakangas, Andres Freund)
       </P
><P
>        The syntax is <TT
CLASS="COMMAND"
>INSERT ... ON CONFLICT DO NOTHING/UPDATE</TT
>.
        This is the Postgres implementation of the popular
        <TT
CLASS="COMMAND"
>UPSERT</TT
> command.
       </P
></LI
><LI
><P
>        Add <TT
CLASS="LITERAL"
>GROUP BY</TT
> analysis features <A
HREF="queries-table-expressions.html#QUERIES-GROUPING-SETS"
><TT
CLASS="LITERAL"
>GROUPING SETS</TT
></A
>,
        <A
HREF="queries-table-expressions.html#QUERIES-GROUPING-SETS"
><TT
CLASS="LITERAL"
>CUBE</TT
></A
> and
        <A
HREF="queries-table-expressions.html#QUERIES-GROUPING-SETS"
><TT
CLASS="LITERAL"
>ROLLUP</TT
></A
>
        (Andrew Gierth, Atri Sharma)
       </P
></LI
><LI
><P
>        Allow setting multiple target columns in
        an <A
HREF="sql-update.html"
><TT
CLASS="COMMAND"
>UPDATE</TT
></A
> from the result of
        a single sub-SELECT (Tom Lane)
       </P
><P
>        This is accomplished using the syntax <TT
CLASS="COMMAND"
>UPDATE tab SET
        (col1, col2, ...) = (SELECT ...)</TT
>.
       </P
></LI
><LI
><P
>        Add <A
HREF="sql-select.html"
><TT
CLASS="COMMAND"
>SELECT</TT
></A
> option
        <TT
CLASS="LITERAL"
>SKIP LOCKED</TT
> to skip locked rows (Thomas Munro)
       </P
><P
>        This does not throw an error for locked rows like
        <TT
CLASS="LITERAL"
>NOWAIT</TT
> does.
       </P
></LI
><LI
><P
>        Add <A
HREF="sql-select.html"
><TT
CLASS="COMMAND"
>SELECT</TT
></A
> option
        <TT
CLASS="LITERAL"
>TABLESAMPLE</TT
> to return a subset of a table (Petr
        Jel&iacute;nek)
       </P
><P
>        This feature supports the SQL-standard table sampling methods.
        In addition, there are provisions
        for <A
HREF="tablesample-method.html"
>user-defined
        table sampling methods</A
>.
       </P
></LI
><LI
><P
>        Suggest possible matches for mistyped column names (Peter
        Geoghegan, Robert Haas)
       </P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN136735"
>E.26.3.4. Utility Commands</A
></H3
><P
></P
><UL
><LI
><P
>        Add more details about sort ordering in <A
HREF="sql-explain.html"
><TT
CLASS="COMMAND"
>EXPLAIN</TT
></A
> output (Marius Timmer,
        Lukas Kreft, Arne Scheffer)
       </P
><P
>        Details include <TT
CLASS="LITERAL"
>COLLATE</TT
>, <TT
CLASS="LITERAL"
>DESC</TT
>,
        <TT
CLASS="LITERAL"
>USING</TT
>, and <TT
CLASS="LITERAL"
>NULLS FIRST</TT
><TT
CLASS="LITERAL"
>/LAST</TT
>.
       </P
></LI
><LI
><P
>        Make <A
HREF="sql-vacuum.html"
><TT
CLASS="COMMAND"
>VACUUM</TT
></A
> log the
        number of pages skipped due to pins (Jim Nasby)
       </P
></LI
><LI
><P
>        Make <A
HREF="sql-truncate.html"
><TT
CLASS="COMMAND"
>TRUNCATE</TT
></A
> properly
        update the <TT
CLASS="LITERAL"
>pg_stat</TT
>* tuple counters (Alexander Shulgin)
       </P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN136757"
>E.26.3.4.1. <A
HREF="sql-reindex.html"
>REINDEX</A
></A
></H4
><P
></P
><UL
><LI
><P
>        Allow <TT
CLASS="COMMAND"
>REINDEX</TT
> to reindex an entire schema using the
        <TT
CLASS="LITERAL"
>SCHEMA</TT
> option (Sawada Masahiko)
       </P
></LI
><LI
><P
>        Add <TT
CLASS="LITERAL"
>VERBOSE</TT
> option to <TT
CLASS="COMMAND"
>REINDEX</TT
> (Sawada
        Masahiko)
       </P
></LI
><LI
><P
>        Prevent <TT
CLASS="COMMAND"
>REINDEX DATABASE</TT
> and <TT
CLASS="COMMAND"
>SCHEMA</TT
>
        from outputting object names, unless <TT
CLASS="LITERAL"
>VERBOSE</TT
> is used
        (Simon Riggs)
       </P
></LI
><LI
><P
>        Remove obsolete <TT
CLASS="LITERAL"
>FORCE</TT
> option from <TT
CLASS="COMMAND"
>REINDEX</TT
>
        (Fujii Masao)
       </P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN136778"
>E.26.3.5. Object Manipulation</A
></H3
><P
></P
><UL
><LI
><P
>        Add  <A
HREF="ddl-rowsecurity.html"
>row-level security control</A
>
        (Craig Ringer, KaiGai Kohei, Adam Brightwell, Dean Rasheed,
        Stephen Frost)
       </P
><P
>        This feature allows row-by-row control over which users can add,
        modify, or even see rows in a table.  This is controlled by new
        commands <A
HREF="sql-createpolicy.html"
><TT
CLASS="COMMAND"
>CREATE</TT
></A
>/<A
HREF="sql-alterpolicy.html"
><TT
CLASS="COMMAND"
>ALTER</TT
></A
>/<A
HREF="sql-droppolicy.html"
><TT
CLASS="COMMAND"
>DROP POLICY</TT
></A
> and <A
HREF="sql-altertable.html"
><TT
CLASS="COMMAND"
>ALTER TABLE ... ENABLE/DISABLE
        ROW SECURITY</TT
></A
>.
       </P
></LI
><LI
><P
>        Allow changing of the <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
>
        logging status of a table after creation with <A
HREF="sql-altertable.html"
><TT
CLASS="COMMAND"
>ALTER TABLE ... SET LOGGED /
        UNLOGGED</TT
></A
> (Fabr&iacute;zio de Royes Mello)
       </P
></LI
><LI
><P
>        Add <TT
CLASS="LITERAL"
>IF NOT EXISTS</TT
> clause to <A
HREF="sql-createtableas.html"
><TT
CLASS="COMMAND"
>CREATE TABLE AS</TT
></A
>,
        <A
HREF="sql-createindex.html"
><TT
CLASS="COMMAND"
>CREATE INDEX</TT
></A
>,
        <A
HREF="sql-createsequence.html"
><TT
CLASS="COMMAND"
>CREATE SEQUENCE</TT
></A
>,
        and <A
HREF="sql-creatematerializedview.html"
><TT
CLASS="COMMAND"
>CREATE
        MATERIALIZED VIEW</TT
></A
> (Fabr&iacute;zio de Royes Mello)
       </P
></LI
><LI
><P
>        Add support for <TT
CLASS="LITERAL"
>IF EXISTS</TT
> to <A
HREF="sql-altertable.html"
><TT
CLASS="COMMAND"
>ALTER TABLE ... RENAME
        CONSTRAINT</TT
></A
> (Bruce Momjian)
       </P
></LI
><LI
><P
>        Allow some DDL commands to accept <TT
CLASS="LITERAL"
>CURRENT_USER</TT
>
        or <TT
CLASS="LITERAL"
>SESSION_USER</TT
>, meaning the current user or session
        user, in place of a specific user name (Kyotaro Horiguchi,
        &Aacute;lvaro Herrera)
       </P
><P
>        This feature is now supported in
        <A
HREF="sql-alteruser.html"
>ALTER USER</A
>, <A
HREF="sql-altergroup.html"
>ALTER GROUP</A
>,
        <A
HREF="sql-alterrole.html"
>ALTER ROLE</A
>, <A
HREF="sql-grant.html"
>GRANT</A
>,
        and <TT
CLASS="COMMAND"
>ALTER <TT
CLASS="REPLACEABLE"
><I
>object</I
></TT
> OWNER TO</TT
> commands.
       </P
></LI
><LI
><P
>        Support comments on <A
HREF="sql-createdomain.html"
>domain
        constraints</A
> (&Aacute;lvaro Herrera)
       </P
></LI
><LI
><P
>        Reduce lock levels of some create and alter trigger and foreign
        key commands (Simon Riggs, Andreas Karlsson)
       </P
></LI
><LI
><P
>        Allow <A
HREF="sql-lock.html"
><TT
CLASS="COMMAND"
>LOCK TABLE ... ROW EXCLUSIVE
        MODE</TT
></A
> for those with <TT
CLASS="COMMAND"
>INSERT</TT
> privileges on the
        target table (Stephen Frost)
       </P
><P
>        Previously this command required <TT
CLASS="COMMAND"
>UPDATE</TT
>, <TT
CLASS="COMMAND"
>DELETE</TT
>,
        or <TT
CLASS="COMMAND"
>TRUNCATE</TT
> privileges.
       </P
></LI
><LI
><P
>        Apply table and domain <TT
CLASS="LITERAL"
>CHECK</TT
> constraints in order by name
        (Tom Lane)
       </P
><P
>        The previous ordering was indeterminate.
       </P
></LI
><LI
><P
>        Allow <A
HREF="sql-createdatabase.html"
><TT
CLASS="COMMAND"
>CREATE</TT
></A
>/<A
HREF="sql-alterdatabase.html"
><TT
CLASS="COMMAND"
>ALTER DATABASE</TT
></A
>
        to manipulate <TT
CLASS="STRUCTNAME"
>datistemplate</TT
> and
        <TT
CLASS="STRUCTNAME"
>datallowconn</TT
> (Vik Fearing)
       </P
><P
>        This allows these per-database settings to be
        changed without manually modifying the <A
HREF="catalog-pg-database.html"
><TT
CLASS="STRUCTNAME"
>pg_database</TT
></A
>
        system catalog.
       </P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN136854"
>E.26.3.5.1. <A
HREF="ddl-foreign-data.html"
>Foreign Tables</A
></A
></H4
><P
></P
><UL
><LI
><P
>        Add support for <A
HREF="sql-importforeignschema.html"
>IMPORT FOREIGN SCHEMA</A
>
        (Ronan Dunklau, Michael Paquier, Tom Lane)
       </P
><P
>        This command allows automatic creation of local foreign tables
        that match the structure of existing tables on a remote server.
       </P
></LI
><LI
><P
>        Allow <TT
CLASS="LITERAL"
>CHECK</TT
> constraints to be placed on foreign tables
        (Shigeru Hanada, Etsuro Fujita)
       </P
><P
>        Such constraints are assumed to be enforced on the remote server,
        and are not enforced locally.  However, they are assumed to hold for
        purposes of query optimization, such
        as <A
HREF="ddl-partitioning.html#DDL-PARTITIONING-CONSTRAINT-EXCLUSION"
>constraint
        exclusion</A
>.
       </P
></LI
><LI
><P
>        Allow foreign tables to participate in inheritance (Shigeru Hanada,
        Etsuro Fujita)
       </P
><P
>        To let this work naturally, foreign tables are now allowed to have
        check constraints marked as not valid, and to set storage
        and <TT
CLASS="TYPE"
>OID</TT
> characteristics, even though these operations are
        effectively no-ops for a foreign table.
       </P
></LI
><LI
><P
>        Allow foreign data wrappers and custom scans to implement join
        pushdown (KaiGai Kohei)
       </P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN136873"
>E.26.3.5.2. <A
HREF="event-triggers.html"
>Event Triggers</A
></A
></H4
><P
></P
><UL
><LI
><P
>         Whenever a <TT
CLASS="LITERAL"
>ddl_command_end</TT
> event trigger is installed,
         capture details of <ACRONYM
CLASS="ACRONYM"
>DDL</ACRONYM
> activity for it to inspect
         (&Aacute;lvaro Herrera)
        </P
><P
>         This information is available through a set-returning function <A
HREF="functions-event-triggers.html#PG-EVENT-TRIGGER-DDL-COMMAND-END-FUNCTIONS"
><CODE
CLASS="FUNCTION"
>pg_event_trigger_ddl_commands()</CODE
></A
>,
         or by inspection of C data structures if that function doesn't
         provide enough detail.
        </P
></LI
><LI
><P
>         Allow event triggers on table rewrites caused by <A
HREF="sql-altertable.html"
><TT
CLASS="COMMAND"
>ALTER TABLE</TT
></A
> (Dimitri
         Fontaine)
        </P
></LI
><LI
><P
>         Add event trigger support for database-level <A
HREF="sql-comment.html"
><TT
CLASS="COMMAND"
>COMMENT</TT
></A
>, <A
HREF="sql-security-label.html"
><TT
CLASS="COMMAND"
>SECURITY LABEL</TT
></A
>,
         and <A
HREF="sql-grant.html"
><TT
CLASS="COMMAND"
>GRANT</TT
></A
>/<A
HREF="sql-revoke.html"
><TT
CLASS="COMMAND"
>REVOKE</TT
></A
> (&Aacute;lvaro Herrera)
        </P
></LI
><LI
><P
>         Add columns to the output of <A
HREF="functions-event-triggers.html#PG-EVENT-TRIGGER-SQL-DROP-FUNCTIONS"
><CODE
CLASS="FUNCTION"
>pg_event_trigger_dropped_objects</CODE
></A
>
         (&Aacute;lvaro Herrera)
        </P
><P
>         This allows simpler processing of delete operations.
        </P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN136903"
>E.26.3.6. Data Types</A
></H3
><P
></P
><UL
><LI
><P
>       Allow the <A
HREF="datatype-xml.html"
><TT
CLASS="TYPE"
>xml</TT
></A
> data type
       to accept empty or all-whitespace content values (Peter Eisentraut)
      </P
><P
>       This is required by the <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
>/<ACRONYM
CLASS="ACRONYM"
>XML</ACRONYM
>
       specification.
      </P
></LI
><LI
><P
>       Allow <A
HREF="datatype-net-types.html#DATATYPE-MACADDR"
><TT
CLASS="TYPE"
>macaddr</TT
></A
> input
       using the format <TT
CLASS="LITERAL"
>xxxx-xxxx-xxxx</TT
> (Herwin Weststrate)
      </P
></LI
><LI
><P
>       Disallow non-SQL-standard syntax for <A
HREF="datatype-datetime.html#DATATYPE-INTERVAL-INPUT"
><TT
CLASS="TYPE"
>interval</TT
></A
> with
       both precision and field specifications (Bruce Momjian)
      </P
><P
>       Per the standard, such type specifications should be written as,
       for example, <TT
CLASS="LITERAL"
>INTERVAL MINUTE TO SECOND(2)</TT
>.
       <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> formerly allowed this to be written as
       <TT
CLASS="LITERAL"
>INTERVAL(2) MINUTE TO SECOND</TT
>, but it must now be
       written in the standard way.
      </P
></LI
><LI
><P
>       Add selectivity estimators for <A
HREF="datatype-net-types.html#DATATYPE-INET"
><TT
CLASS="TYPE"
>inet</TT
></A
>/<A
HREF="datatype-net-types.html#DATATYPE-CIDR"
><TT
CLASS="TYPE"
>cidr</TT
></A
> operators and improve
       estimators for text search functions (Emre Hasegeli, Tom Lane)
      </P
></LI
><LI
><P
>       Add data
       types <A
HREF="datatype-oid.html#DATATYPE-OID-TABLE"
><TT
CLASS="TYPE"
>regrole</TT
></A
>
       and <A
HREF="datatype-oid.html#DATATYPE-OID-TABLE"
><TT
CLASS="TYPE"
>regnamespace</TT
></A
>
       to simplify entering and pretty-printing the <TT
CLASS="TYPE"
>OID</TT
> of a role
       or namespace (Kyotaro Horiguchi)
      </P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN136939"
>E.26.3.6.1. <A
HREF="datatype-json.html"
><ACRONYM
CLASS="ACRONYM"
>JSON</ACRONYM
></A
></A
></H4
><P
></P
><UL
><LI
><P
>        Add <TT
CLASS="TYPE"
>jsonb</TT
> functions <A
HREF="functions-json.html#FUNCTIONS-JSON-PROCESSING-TABLE"
><CODE
CLASS="FUNCTION"
>jsonb_set()</CODE
></A
>
        and <A
HREF="functions-json.html#FUNCTIONS-JSON-PROCESSING-TABLE"
><CODE
CLASS="FUNCTION"
>jsonb_pretty()</CODE
></A
>
        (Dmitry Dolgov, Andrew Dunstan, Petr Jel&iacute;nek)
       </P
></LI
><LI
><P
>        Add <TT
CLASS="TYPE"
>jsonb</TT
> generator functions <A
HREF="functions-json.html#FUNCTIONS-JSON-CREATION-TABLE"
><CODE
CLASS="FUNCTION"
>to_jsonb()</CODE
></A
>,
        <A
HREF="functions-json.html#FUNCTIONS-JSON-CREATION-TABLE"
><CODE
CLASS="FUNCTION"
>jsonb_object()</CODE
></A
>,
        <A
HREF="functions-json.html#FUNCTIONS-JSON-CREATION-TABLE"
><CODE
CLASS="FUNCTION"
>jsonb_build_object()</CODE
></A
>,
        <A
HREF="functions-json.html#FUNCTIONS-JSON-CREATION-TABLE"
><CODE
CLASS="FUNCTION"
>jsonb_build_array()</CODE
></A
>,
        <A
HREF="functions-aggregate.html#FUNCTIONS-AGGREGATE-TABLE"
><CODE
CLASS="FUNCTION"
>jsonb_agg()</CODE
></A
>,
        and <A
HREF="functions-aggregate.html#FUNCTIONS-AGGREGATE-TABLE"
><CODE
CLASS="FUNCTION"
>jsonb_object_agg()</CODE
></A
>
        (Andrew Dunstan)
       </P
><P
>        Equivalent functions already existed for type <TT
CLASS="TYPE"
>json</TT
>.
       </P
></LI
><LI
><P
>        Reduce casting requirements to/from <A
HREF="datatype-json.html"
><TT
CLASS="TYPE"
>json</TT
></A
> and <A
HREF="datatype-json.html"
><TT
CLASS="TYPE"
>jsonb</TT
></A
> (Tom Lane)
       </P
></LI
><LI
><P
>        Allow <TT
CLASS="TYPE"
>text</TT
>, <TT
CLASS="TYPE"
>text</TT
> array, and <TT
CLASS="TYPE"
>integer</TT
>
        values to be <A
HREF="functions-json.html#FUNCTIONS-JSONB-OP-TABLE"
>subtracted</A
>
        from <TT
CLASS="TYPE"
>jsonb</TT
> documents (Dmitry Dolgov, Andrew Dunstan)
       </P
></LI
><LI
><P
>        Add <TT
CLASS="TYPE"
>jsonb</TT
> <A
HREF="functions-json.html#FUNCTIONS-JSONB-OP-TABLE"
><TT
CLASS="LITERAL"
>||</TT
></A
> operator
        (Dmitry Dolgov, Andrew Dunstan)
       </P
></LI
><LI
><P
>        Add <A
HREF="functions-json.html#FUNCTIONS-JSON-PROCESSING-TABLE"
><CODE
CLASS="FUNCTION"
>json_strip_nulls()</CODE
></A
>
        and <A
HREF="functions-json.html#FUNCTIONS-JSON-PROCESSING-TABLE"
><CODE
CLASS="FUNCTION"
>jsonb_strip_nulls()</CODE
></A
>
        functions to remove JSON null values from documents
        (Andrew Dunstan)
       </P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN136992"
>E.26.3.7. Functions</A
></H3
><P
></P
><UL
><LI
><P
>        Add <A
HREF="functions-srf.html"
><CODE
CLASS="FUNCTION"
>generate_series()</CODE
></A
>
        for <TT
CLASS="TYPE"
>numeric</TT
> values (Plato Malugin)
       </P
></LI
><LI
><P
>        Allow <A
HREF="functions-aggregate.html#FUNCTIONS-AGGREGATE-TABLE"
><CODE
CLASS="FUNCTION"
>array_agg()</CODE
></A
> and
        <CODE
CLASS="FUNCTION"
>ARRAY()</CODE
> to take arrays as inputs (Ali Akbar, Tom Lane)
       </P
></LI
><LI
><P
>        Add functions <A
HREF="functions-array.html#ARRAY-FUNCTIONS-TABLE"
><CODE
CLASS="FUNCTION"
>array_position()</CODE
></A
>
        and <A
HREF="functions-array.html#ARRAY-FUNCTIONS-TABLE"
><CODE
CLASS="FUNCTION"
>array_positions()</CODE
></A
>
        to return subscripts of array values (Pavel Stehule)
       </P
></LI
><LI
><P
>        Add a <TT
CLASS="TYPE"
>point</TT
>-to-<TT
CLASS="TYPE"
>polygon</TT
> distance operator
        <A
HREF="functions-geometry.html#FUNCTIONS-GEOMETRY-OP-TABLE"
>&lt;-&gt;</A
>
        (Alexander Korotkov)
       </P
></LI
><LI
><P
>        Allow multibyte characters as escapes in <A
HREF="functions-matching.html#FUNCTIONS-SIMILARTO-REGEXP"
><TT
CLASS="LITERAL"
>SIMILAR TO</TT
></A
>
        and <A
HREF="functions-string.html#FUNCTIONS-STRING-SQL"
><TT
CLASS="LITERAL"
>SUBSTRING</TT
></A
>
        (Jeff Davis)
       </P
><P
>        Previously, only a single-byte character was allowed as an escape.
       </P
></LI
><LI
><P
>        Add a <A
HREF="functions-math.html#FUNCTIONS-MATH-FUNC-TABLE"
><CODE
CLASS="FUNCTION"
>width_bucket()</CODE
></A
>
        variant that supports any sortable data type and non-uniform bucket
        widths (Petr Jel&iacute;nek)
       </P
></LI
><LI
><P
>        Add an optional <TT
CLASS="REPLACEABLE"
><I
>missing_ok</I
></TT
> argument to <A
HREF="functions-admin.html#FUNCTIONS-ADMIN-GENFILE-TABLE"
><CODE
CLASS="FUNCTION"
>pg_read_file()</CODE
></A
>
        and related functions (Michael Paquier, Heikki Linnakangas)
       </P
></LI
><LI
><P
>        Allow <A
HREF="sql-syntax-calling-funcs.html"
><TT
CLASS="LITERAL"
>=&gt;</TT
></A
>
        to specify named parameters in function calls (Pavel Stehule)
       </P
><P
>        Previously only <TT
CLASS="LITERAL"
>:=</TT
> could be used.  This requires removing
        the possibility for <TT
CLASS="LITERAL"
>=&gt;</TT
> to be a user-defined operator.
        Creation of user-defined <TT
CLASS="LITERAL"
>=&gt;</TT
> operators has been issuing
        warnings since PostgreSQL 9.0.
       </P
></LI
><LI
><P
>        Add <ACRONYM
CLASS="ACRONYM"
>POSIX</ACRONYM
>-compliant rounding for platforms that use
        PostgreSQL-supplied rounding functions (Pedro Gimeno Fortea)
       </P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN137043"
>E.26.3.7.1. System Information Functions and Views</A
></H4
><P
></P
><UL
><LI
><P
>        Add function <A
HREF="functions-info.html#FUNCTIONS-INFO-OBJECT-TABLE"
><CODE
CLASS="FUNCTION"
>pg_get_object_address()</CODE
></A
>
        to return <TT
CLASS="TYPE"
>OID</TT
>s that uniquely
        identify an object, and function <A
HREF="functions-info.html#FUNCTIONS-INFO-OBJECT-TABLE"
><CODE
CLASS="FUNCTION"
>pg_identify_object_as_address()</CODE
></A
>
        to return object information based on <TT
CLASS="TYPE"
>OID</TT
>s (&Aacute;lvaro
        Herrera)
       </P
></LI
><LI
><P
>        Loosen security checks for viewing queries in <A
HREF="monitoring-stats.html#PG-STAT-ACTIVITY-VIEW"
><TT
CLASS="STRUCTNAME"
>pg_stat_activity</TT
></A
>,
        executing <A
HREF="functions-admin.html#FUNCTIONS-ADMIN-SIGNAL-TABLE"
><CODE
CLASS="FUNCTION"
>pg_cancel_backend()</CODE
></A
>,
        and executing <A
HREF="functions-admin.html#FUNCTIONS-ADMIN-SIGNAL-TABLE"
><CODE
CLASS="FUNCTION"
>pg_terminate_backend()</CODE
></A
>
        (Stephen Frost)
       </P
><P
>        Previously, only the specific role owning the target session could
        perform these operations; now membership in that role is sufficient.
       </P
></LI
><LI
><P
>        Add <A
HREF="monitoring-stats.html#MONITORING-STATS-FUNCS-TABLE"
><CODE
CLASS="FUNCTION"
>pg_stat_get_snapshot_timestamp()</CODE
></A
>
        to output the time stamp of the statistics snapshot (Matt Kelly)
       </P
><P
>        This represents the last time the snapshot file was written to
        the file system.
       </P
></LI
><LI
><P
>        Add <A
HREF="routine-vacuuming.html#VACUUM-FOR-MULTIXACT-WRAPAROUND"
><CODE
CLASS="FUNCTION"
>mxid_age()</CODE
></A
>
        to compute multi-xid age (Bruce Momjian)
       </P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN137072"
>E.26.3.7.2. Aggregates</A
></H4
><P
></P
><UL
><LI
><P
>        Add <CODE
CLASS="FUNCTION"
>min()</CODE
>/<CODE
CLASS="FUNCTION"
>max()</CODE
> aggregates
        for <A
HREF="datatype-net-types.html#DATATYPE-INET"
><TT
CLASS="TYPE"
>inet</TT
></A
>/<A
HREF="datatype-net-types.html#DATATYPE-CIDR"
><TT
CLASS="TYPE"
>cidr</TT
></A
> data types (Haribabu
        Kommi)
       </P
></LI
><LI
><P
>        Use 128-bit integers, where supported, as accumulators for some
        aggregate functions (Andreas Karlsson)
       </P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN137085"
>E.26.3.8. Server-Side Languages</A
></H3
><P
></P
><UL
><LI
><P
>       Improve support for composite types in <A
HREF="plpython.html"
><SPAN
CLASS="APPLICATION"
>PL/Python</SPAN
></A
> (Ed Behn, Ronan
       Dunklau)
      </P
><P
>       This allows <SPAN
CLASS="APPLICATION"
>PL/Python</SPAN
> functions to return arrays
       of composite types.
      </P
></LI
><LI
><P
>       Reduce lossiness of <A
HREF="plpython.html"
><SPAN
CLASS="APPLICATION"
>PL/Python</SPAN
></A
> floating-point value
       conversions (Marko Kreen)
      </P
></LI
><LI
><P
>       Allow specification of conversion routines between <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
>
       data types and data types of procedural languages (Peter Eisentraut)
      </P
><P
>       This change adds new commands <A
HREF="sql-createtransform.html"
><TT
CLASS="COMMAND"
>CREATE</TT
></A
>/<A
HREF="sql-droptransform.html"
><TT
CLASS="COMMAND"
>DROP TRANSFORM</TT
></A
>.
       This also adds optional transformations between the <A
HREF="hstore.html"
><SPAN
CLASS="APPLICATION"
>hstore</SPAN
></A
> and <A
HREF="ltree.html"
><SPAN
CLASS="APPLICATION"
>ltree</SPAN
></A
> types to/from <A
HREF="plperl.html"
><SPAN
CLASS="APPLICATION"
>PL/Perl</SPAN
></A
> and <A
HREF="plpython.html"
><SPAN
CLASS="APPLICATION"
>PL/Python</SPAN
></A
>.
      </P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN137114"
>E.26.3.8.1. <A
HREF="plpgsql.html"
>PL/pgSQL</A
> Server-Side Language</A
></H4
><P
></P
><UL
><LI
><P
>       Improve <A
HREF="plpgsql.html"
><SPAN
CLASS="APPLICATION"
>PL/pgSQL</SPAN
></A
> array
       performance (Tom Lane)
      </P
></LI
><LI
><P
>       Add an <A
HREF="plpgsql-errors-and-messages.html#PLPGSQL-STATEMENTS-ASSERT"
><TT
CLASS="COMMAND"
>ASSERT</TT
></A
>
       statement in <SPAN
CLASS="APPLICATION"
>PL/pgSQL</SPAN
> (Pavel Stehule)
      </P
></LI
><LI
><P
>       Allow more <A
HREF="plpgsql.html"
><SPAN
CLASS="APPLICATION"
>PL/pgSQL</SPAN
></A
>
       keywords to be used as identifiers (Tom Lane)
      </P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN137131"
>E.26.3.9. Client Applications</A
></H3
><P
></P
><UL
><LI
><P
>       Move <A
HREF="pgarchivecleanup.html"
><SPAN
CLASS="APPLICATION"
>pg_archivecleanup</SPAN
></A
>,
       <A
HREF="pgtestfsync.html"
><SPAN
CLASS="APPLICATION"
>pg_test_fsync</SPAN
></A
>,
       <A
HREF="pgtesttiming.html"
><SPAN
CLASS="APPLICATION"
>pg_test_timing</SPAN
></A
>,
       and <A
HREF="pgxlogdump.html"
><SPAN
CLASS="APPLICATION"
>pg_xlogdump</SPAN
></A
>
       from <TT
CLASS="FILENAME"
>contrib</TT
> to <TT
CLASS="FILENAME"
>src/bin</TT
> (Peter Eisentraut)
      </P
><P
>       This should result in these programs being installed by default in
       most installations.
      </P
></LI
><LI
><P
>       Add <A
HREF="app-pgrewind.html"
><SPAN
CLASS="APPLICATION"
>pg_rewind</SPAN
></A
>,
       which allows re-synchronizing a master server after failback
       (Heikki Linnakangas)
      </P
></LI
><LI
><P
>       Allow <A
HREF="app-pgreceivexlog.html"
><SPAN
CLASS="APPLICATION"
>pg_receivexlog</SPAN
></A
>
       to manage physical replication slots (Michael Paquier)
      </P
><P
>       This is controlled via new <TT
CLASS="OPTION"
>--create-slot</TT
> and
       <TT
CLASS="OPTION"
>--drop-slot</TT
> options.
      </P
></LI
><LI
><P
>       Allow <A
HREF="app-pgreceivexlog.html"
><SPAN
CLASS="APPLICATION"
>pg_receivexlog</SPAN
></A
>
       to synchronously flush <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> to storage using new
       <TT
CLASS="OPTION"
>--synchronous</TT
> option (Furuya Osamu, Fujii Masao)
      </P
><P
>       Without this, <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> files are fsync'ed only on close.
      </P
></LI
><LI
><P
>       Allow <A
HREF="app-vacuumdb.html"
><SPAN
CLASS="APPLICATION"
>vacuumdb</SPAN
></A
> to
       vacuum in parallel using new <TT
CLASS="OPTION"
>--jobs</TT
> option (Dilip Kumar)
      </P
></LI
><LI
><P
>       In <A
HREF="app-vacuumdb.html"
><SPAN
CLASS="APPLICATION"
>vacuumdb</SPAN
></A
>, do not
       prompt for the same password repeatedly when multiple connections
       are necessary (Haribabu Kommi, Michael Paquier)
      </P
></LI
><LI
><P
>       Add <TT
CLASS="OPTION"
>--verbose</TT
> option to <A
HREF="app-reindexdb.html"
><SPAN
CLASS="APPLICATION"
>reindexdb</SPAN
></A
> (Sawada
       Masahiko)
      </P
></LI
><LI
><P
>       Make <A
HREF="app-pgbasebackup.html"
><SPAN
CLASS="APPLICATION"
>pg_basebackup</SPAN
></A
>
       use a tablespace mapping file when using <SPAN
CLASS="APPLICATION"
>tar</SPAN
> format,
       to support symbolic links and file paths of 100+ characters in length
       on <SPAN
CLASS="SYSTEMITEM"
>MS Windows</SPAN
> (Amit Kapila)
      </P
></LI
><LI
><P
>       Add <A
HREF="pgxlogdump.html"
><SPAN
CLASS="APPLICATION"
>pg_xlogdump</SPAN
></A
> option
       <TT
CLASS="OPTION"
>--stats</TT
> to display summary statistics (Abhijit Menon-Sen)
      </P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN137191"
>E.26.3.9.1. <A
HREF="app-psql.html"
><SPAN
CLASS="APPLICATION"
>psql</SPAN
></A
></A
></H4
><P
></P
><UL
><LI
><P
>         Allow <SPAN
CLASS="APPLICATION"
>psql</SPAN
> to produce AsciiDoc output (Szymon Guz)
        </P
></LI
><LI
><P
>         Add an <TT
CLASS="LITERAL"
>errors</TT
> mode that displays only failed commands
         to <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s <TT
CLASS="VARNAME"
>ECHO</TT
> variable
         (Pavel Stehule)
        </P
><P
>         This behavior can also be selected with <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s
         <TT
CLASS="OPTION"
>-b</TT
> option.
        </P
></LI
><LI
><P
>         Provide separate column, header, and border linestyle control
         in <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s unicode linestyle (Pavel Stehule)
        </P
><P
>         Single or double lines are supported; the default is
         <TT
CLASS="LITERAL"
>single</TT
>.
        </P
></LI
><LI
><P
>         Add new option <TT
CLASS="LITERAL"
>%l</TT
> in <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s <A
HREF="app-psql.html#APP-PSQL-VARIABLES"
><TT
CLASS="ENVAR"
>PROMPT</TT
></A
> variables
         to display the current multiline statement line number
         (Sawada Masahiko)
        </P
></LI
><LI
><P
>         Add <TT
CLASS="LITERAL"
>\pset</TT
> option <A
HREF="app-psql.html#APP-PSQL-META-COMMANDS"
><TT
CLASS="VARNAME"
>pager_min_lines</TT
></A
>
         to control pager invocation (Andrew Dunstan)
        </P
></LI
><LI
><P
>         Improve <SPAN
CLASS="APPLICATION"
>psql</SPAN
> line counting used when deciding
         to invoke the pager (Andrew Dunstan)
        </P
></LI
><LI
><P
>         <SPAN
CLASS="APPLICATION"
>psql</SPAN
> now fails if the file specified by
         an <TT
CLASS="OPTION"
>--output</TT
> or <TT
CLASS="OPTION"
>--log-file</TT
> switch cannot be
         written (Tom Lane, Daniel V&eacute;rit&eacute;)
        </P
><P
>         Previously, it effectively ignored the switch in such cases.
        </P
></LI
><LI
><P
>         Add <SPAN
CLASS="APPLICATION"
>psql</SPAN
> tab completion when setting the
         <A
HREF="runtime-config-client.html#GUC-SEARCH-PATH"
>search_path</A
> variable (Jeff Janes)
        </P
><P
>         Currently only the first schema can be tab-completed.
        </P
></LI
><LI
><P
>         Improve <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s tab completion for triggers and rules
         (Andreas Karlsson)
        </P
></LI
></UL
><DIV
CLASS="SECT5"
><H5
CLASS="SECT5"
><A
NAME="AEN137239"
>E.26.3.9.1.1. <A
HREF="app-psql.html#APP-PSQL-META-COMMANDS"
>Backslash Commands</A
></A
></H5
><P
></P
><UL
><LI
><P
>         Add <SPAN
CLASS="APPLICATION"
>psql</SPAN
> <TT
CLASS="COMMAND"
>\?</TT
> help sections
         <TT
CLASS="LITERAL"
>variables</TT
> and <TT
CLASS="LITERAL"
>options</TT
> (Pavel Stehule)
        </P
><P
>         <TT
CLASS="LITERAL"
>\? variables</TT
> shows <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s special
         variables and <TT
CLASS="LITERAL"
>\? options</TT
> shows the command-line options.
         <TT
CLASS="COMMAND"
>\? commands</TT
> shows the meta-commands, which is the
         traditional output and remains the default.  These help displays
         can also be obtained with the command-line
         option <TT
CLASS="LITERAL"
>--help=<TT
CLASS="REPLACEABLE"
><I
>section</I
></TT
></TT
>.
        </P
></LI
><LI
><P
>         Show tablespace size in <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s <TT
CLASS="LITERAL"
>\db+</TT
>
         (Fabr&iacute;zio de Royes Mello)
        </P
></LI
><LI
><P
>         Show data type owners in <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s <TT
CLASS="LITERAL"
>\dT+</TT
>
         (Magnus Hagander)
        </P
></LI
><LI
><P
>         Allow <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s <TT
CLASS="COMMAND"
>\watch</TT
> to output
         <TT
CLASS="COMMAND"
>\timing</TT
> information (Fujii Masao)
        </P
><P
>         Also prevent <TT
CLASS="OPTION"
>--echo-hidden</TT
> from echoing
         <TT
CLASS="COMMAND"
>\watch</TT
> queries, since that is generally unwanted.
        </P
></LI
><LI
><P
>         Make <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s <TT
CLASS="LITERAL"
>\sf</TT
> and <TT
CLASS="LITERAL"
>\ef</TT
>
         commands honor <TT
CLASS="ENVAR"
>ECHO_HIDDEN</TT
> (Andrew Dunstan)
        </P
></LI
><LI
><P
>         Improve <SPAN
CLASS="APPLICATION"
>psql</SPAN
> tab completion for <TT
CLASS="COMMAND"
>\set</TT
>,
         <TT
CLASS="COMMAND"
>\unset</TT
>, and <TT
CLASS="LITERAL"
>:variable</TT
> names (Pavel
         Stehule)
        </P
></LI
><LI
><P
>         Allow tab completion of role names
         in <SPAN
CLASS="APPLICATION"
>psql</SPAN
> <TT
CLASS="LITERAL"
>\c</TT
> commands (Ian Barwick)
        </P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN137288"
>E.26.3.9.2. <A
HREF="app-pgdump.html"
>pg_dump</A
></A
></H4
><P
></P
><UL
><LI
><P
>        Allow <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> to share a snapshot taken by another
        session using <TT
CLASS="OPTION"
>--snapshot</TT
> (Simon Riggs, Michael Paquier)
       </P
><P
>        The remote snapshot must have been exported by
        <CODE
CLASS="FUNCTION"
>pg_export_snapshot()</CODE
> or logical replication slot
        creation.  This can be used to share a consistent snapshot
        across multiple <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> processes.
       </P
></LI
><LI
><P
>        Support table sizes exceeding 8GB in tar archive format (Tom Lane)
       </P
><P
>        The POSIX standard for tar format does not allow elements of a tar
        archive to exceed 8GB, but most modern implementations of tar
        support an extension that does allow it.  Use the extension format
        when necessary, rather than failing.
       </P
></LI
><LI
><P
>        Make <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> always print the server and
        <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> versions (Jing Wang)
       </P
><P
>        Previously, version information was only printed in
        <TT
CLASS="OPTION"
>--verbose</TT
> mode.
       </P
></LI
><LI
><P
>        Remove the long-ignored <TT
CLASS="OPTION"
>-i</TT
>/<TT
CLASS="OPTION"
>--ignore-version</TT
>
        option from <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
>, <SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
>,
        and <SPAN
CLASS="APPLICATION"
>pg_restore</SPAN
> (Fujii Masao)
       </P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN137315"
>E.26.3.9.3. <A
HREF="app-pg-ctl.html"
><SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
></A
></A
></H4
><P
></P
><UL
><LI
><P
>        Support multiple <SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
> <TT
CLASS="OPTION"
>-o</TT
> options,
        concatenating their values (Bruce Momjian)
       </P
></LI
><LI
><P
>        Allow control of <SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
>'s event source logging
        on <SPAN
CLASS="SYSTEMITEM"
>MS Windows</SPAN
> (MauMau)
       </P
><P
>        This only controls <SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
>, not the server, which
        has separate settings in <TT
CLASS="FILENAME"
>postgresql.conf</TT
>.
       </P
></LI
><LI
><P
>        If the server's listen address is set to a wildcard value
        (<TT
CLASS="LITERAL"
>0.0.0.0</TT
> in IPv4 or <TT
CLASS="LITERAL"
>::</TT
> in IPv6), connect via
        the loopback address rather than trying to use the wildcard address
        literally (Kondo Yuta)
       </P
><P
>        This fix primarily affects Windows, since on other platforms
        <SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
> will prefer to use a Unix-domain socket.
       </P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN137336"
>E.26.3.9.4. <A
HREF="pgupgrade.html"
><SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
></A
></A
></H4
><P
></P
><UL
><LI
><P
>        Move <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> from <TT
CLASS="FILENAME"
>contrib</TT
> to
        <TT
CLASS="FILENAME"
>src/bin</TT
> (Peter Eisentraut)
       </P
><P
>        In connection with this change, the functionality previously
        provided by the <SPAN
CLASS="APPLICATION"
>pg_upgrade_support</SPAN
> module has been
        moved into the core server.
       </P
></LI
><LI
><P
>        Support multiple <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
>
        <TT
CLASS="OPTION"
>-o</TT
>/<TT
CLASS="OPTION"
>-O</TT
> options,
        concatenating their values (Bruce Momjian)
       </P
></LI
><LI
><P
>        Improve database collation comparisons in
        <SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
> (Heikki Linnakangas)
       </P
></LI
><LI
><P
>        Remove support for upgrading from 8.3 clusters (Bruce Momjian)
       </P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN137357"
>E.26.3.9.5. <A
HREF="pgbench.html"
><SPAN
CLASS="APPLICATION"
>pgbench</SPAN
></A
></A
></H4
><P
></P
><UL
><LI
><P
>        Move pgbench from <TT
CLASS="FILENAME"
>contrib</TT
> to <TT
CLASS="FILENAME"
>src/bin</TT
>
        (Peter Eisentraut)
       </P
></LI
><LI
><P
>        Fix calculation of TPS number <SPAN
CLASS="QUOTE"
>"excluding connections
        establishing"</SPAN
> (Tatsuo Ishii, Fabien Coelho)
       </P
><P
>        The overhead for connection establishment was miscalculated whenever
        the number of pgbench threads was less than the number of client
        connections.  Although this is clearly a bug, we won't back-patch it
        into pre-9.5 branches since it makes TPS numbers not comparable to
        previous results.
       </P
></LI
><LI
><P
>        Allow counting of pgbench transactions that take over a specified
        amount of time (Fabien Coelho)
       </P
><P
>        This is controlled by a new <TT
CLASS="OPTION"
>--latency-limit</TT
> option.
       </P
></LI
><LI
><P
>        Allow pgbench to generate Gaussian/exponential distributions
        using <TT
CLASS="COMMAND"
>\setrandom</TT
> (Kondo Mitsumasa, Fabien Coelho)
       </P
></LI
><LI
><P
>        Allow <SPAN
CLASS="APPLICATION"
>pgbench</SPAN
>'s <TT
CLASS="COMMAND"
>\set</TT
> command to handle
        arithmetic expressions containing more than one operator, and add
        <TT
CLASS="LITERAL"
>%</TT
> (modulo) to the set of operators it supports
        (Robert Haas, Fabien Coelho)
       </P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN137381"
>E.26.3.10. Source Code</A
></H3
><P
></P
><UL
><LI
><P
>        Simplify <A
HREF="wal.html"
><ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
></A
> record format
        (Heikki Linnakangas)
       </P
><P
>        This allows external tools to more easily track what blocks
        are modified.
       </P
></LI
><LI
><P
>        Improve the representation of transaction commit and abort WAL
        records (Andres Freund)
       </P
></LI
><LI
><P
>        Add atomic memory operations <ACRONYM
CLASS="ACRONYM"
>API</ACRONYM
> (Andres Freund)
       </P
></LI
><LI
><P
>        Allow custom path and scan methods (KaiGai Kohei, Tom Lane)
       </P
><P
>        This allows extensions greater control over the optimizer and
        executor.
       </P
></LI
><LI
><P
>        Allow foreign data wrappers to do post-filter locking (Etsuro
        Fujita)
       </P
></LI
><LI
><P
>        Foreign tables can now take part in <TT
CLASS="COMMAND"
>INSERT ... ON CONFLICT
        DO NOTHING</TT
> queries (Peter Geoghegan, Heikki Linnakangas,
        Andres Freund)
       </P
><P
>        Foreign data wrappers must be modified to handle this.
        <TT
CLASS="COMMAND"
>INSERT ... ON CONFLICT DO UPDATE</TT
> is not supported on
        foreign tables.
       </P
></LI
><LI
><P
>        Improve <CODE
CLASS="FUNCTION"
>hash_create()</CODE
>'s API for selecting
        simple-binary-key hash functions (Teodor Sigaev, Tom Lane)
       </P
></LI
><LI
><P
>        Improve parallel execution infrastructure (Robert Haas, Amit
        Kapila, Noah Misch, Rushabh Lathia, Jeevan Chalke)
       </P
></LI
><LI
><P
>        Remove <SPAN
CLASS="PRODUCTNAME"
>Alpha</SPAN
> (<ACRONYM
CLASS="ACRONYM"
>CPU</ACRONYM
>) and <SPAN
CLASS="SYSTEMITEM"
>Tru64</SPAN
> (OS) ports (Andres Freund)
       </P
></LI
><LI
><P
>        Remove swap-byte-based spinlock implementation for
        <ACRONYM
CLASS="ACRONYM"
>ARM</ACRONYM
>v5 and earlier <ACRONYM
CLASS="ACRONYM"
>CPU</ACRONYM
>s (Robert Haas)
       </P
><P
>        <ACRONYM
CLASS="ACRONYM"
>ARM</ACRONYM
>v5's weak memory ordering made this locking
        implementation unsafe.  Spinlock support is still possible on
        newer gcc implementations with atomics support.
       </P
></LI
><LI
><P
>        Generate an error when excessively long (100+ character) file
        paths are written to tar files (Peter Eisentraut)
       </P
><P
>        Tar does not support such overly-long paths.
       </P
></LI
><LI
><P
>        Change index operator class for columns <A
HREF="catalog-pg-seclabel.html"
><TT
CLASS="STRUCTNAME"
>pg_seclabel</TT
></A
>.<TT
CLASS="STRUCTNAME"
>provider</TT
>
        and <A
HREF="catalog-pg-shseclabel.html"
><TT
CLASS="STRUCTNAME"
>pg_shseclabel</TT
></A
>.<TT
CLASS="STRUCTNAME"
>provider</TT
>
        to be <TT
CLASS="LITERAL"
>text_pattern_ops</TT
> (Tom Lane)
       </P
><P
>        This avoids possible problems with these indexes when different
        databases of a cluster have different default collations.
       </P
></LI
><LI
><P
>        Change the spinlock primitives to function as compiler barriers
        (Robert Haas)
       </P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN137435"
>E.26.3.10.1. MS Windows</A
></H4
><P
></P
><UL
><LI
><P
>        Allow higher-precision time stamp resolution on <SPAN
CLASS="SYSTEMITEM"
>Windows 8</SPAN
>, <SPAN
CLASS="SYSTEMITEM"
>Windows
        Server 2012</SPAN
>, and later Windows systems (Craig Ringer)
       </P
></LI
><LI
><P
>        Install shared libraries to <TT
CLASS="FILENAME"
>bin</TT
> in <SPAN
CLASS="SYSTEMITEM"
>MS Windows</SPAN
> (Peter Eisentraut, Michael Paquier)
       </P
></LI
><LI
><P
>        Install <TT
CLASS="FILENAME"
>src/test/modules</TT
> together with
        <TT
CLASS="FILENAME"
>contrib</TT
> on <SPAN
CLASS="PRODUCTNAME"
>MSVC</SPAN
> builds (Michael
        Paquier)
       </P
></LI
><LI
><P
>        Allow <A
HREF="install-procedure.html"
>configure</A
>'s
        <TT
CLASS="OPTION"
>--with-extra-version</TT
> option to be honored by the
        <SPAN
CLASS="PRODUCTNAME"
>MSVC</SPAN
> build (Michael Paquier)
       </P
></LI
><LI
><P
>        Pass <TT
CLASS="ENVAR"
>PGFILEDESC</TT
> into <SPAN
CLASS="PRODUCTNAME"
>MSVC</SPAN
> contrib builds
        (Michael Paquier)
       </P
></LI
><LI
><P
>        Add icons to all <SPAN
CLASS="PRODUCTNAME"
>MSVC</SPAN
>-built binaries and version
        information to all <SPAN
CLASS="SYSTEMITEM"
>MS Windows</SPAN
>
        binaries (Noah Misch)
       </P
><P
>        MinGW already had such icons.
       </P
></LI
><LI
><P
>        Add optional-argument support to the internal
        <CODE
CLASS="FUNCTION"
>getopt_long()</CODE
> implementation (Michael Paquier,
        Andres Freund)
       </P
><P
>        This is used by the <SPAN
CLASS="PRODUCTNAME"
>MSVC</SPAN
> build.
       </P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN137470"
>E.26.3.11. Additional Modules</A
></H3
><P
></P
><UL
><LI
><P
>       Add statistics for minimum, maximum,
       mean, and standard deviation times to <A
HREF="pgstatstatements.html#PGSTATSTATEMENTS-COLUMNS"
><SPAN
CLASS="APPLICATION"
>pg_stat_statements</SPAN
></A
>
       (Mitsumasa Kondo, Andrew Dunstan)
      </P
></LI
><LI
><P
>       Add <A
HREF="pgcrypto.html"
><SPAN
CLASS="APPLICATION"
>pgcrypto</SPAN
></A
> function
       <CODE
CLASS="FUNCTION"
>pgp_armor_headers()</CODE
> to extract <SPAN
CLASS="PRODUCTNAME"
>PGP</SPAN
>
       armor headers (Marko Tiikkaja, Heikki Linnakangas)
      </P
></LI
><LI
><P
>       Allow empty replacement strings in <A
HREF="unaccent.html"
><SPAN
CLASS="APPLICATION"
>unaccent</SPAN
></A
> (Mohammad Alhashash)
      </P
><P
>       This is useful in languages where diacritic signs are represented
       as separate characters.
      </P
></LI
><LI
><P
>       Allow multicharacter source strings in <A
HREF="unaccent.html"
><SPAN
CLASS="APPLICATION"
>unaccent</SPAN
></A
> (Tom Lane)
      </P
><P
>       This could be useful in languages where diacritic signs are
       represented as separate characters.  It also allows more complex
       unaccent dictionaries.
      </P
></LI
><LI
><P
>       Add <TT
CLASS="FILENAME"
>contrib</TT
> modules <A
HREF="tsm-system-rows.html"
><SPAN
CLASS="APPLICATION"
>tsm_system_rows</SPAN
></A
> and
       <A
HREF="tsm-system-time.html"
><SPAN
CLASS="APPLICATION"
>tsm_system_time</SPAN
></A
>
       to allow additional table sampling methods (Petr Jel&iacute;nek)
      </P
></LI
><LI
><P
>       Add <A
HREF="gin.html"
><ACRONYM
CLASS="ACRONYM"
>GIN</ACRONYM
></A
>
       index inspection functions to <A
HREF="pageinspect.html"
><SPAN
CLASS="APPLICATION"
>pageinspect</SPAN
></A
> (Heikki
       Linnakangas, Peter Geoghegan, Michael Paquier)
      </P
></LI
><LI
><P
>       Add information about buffer pins to <A
HREF="pgbuffercache.html"
><SPAN
CLASS="APPLICATION"
>pg_buffercache</SPAN
></A
> display
       (Andres Freund)
      </P
></LI
><LI
><P
>       Allow <A
HREF="pgstattuple.html"
><SPAN
CLASS="APPLICATION"
>pgstattuple</SPAN
></A
>
       to report approximate answers with less overhead using
       <CODE
CLASS="FUNCTION"
>pgstattuple_approx()</CODE
> (Abhijit Menon-Sen)
      </P
></LI
><LI
><P
>       Move <SPAN
CLASS="APPLICATION"
>dummy_seclabel</SPAN
>, <SPAN
CLASS="APPLICATION"
>test_shm_mq</SPAN
>,
       <SPAN
CLASS="APPLICATION"
>test_parser</SPAN
>, and <SPAN
CLASS="APPLICATION"
>worker_spi</SPAN
>
       from <TT
CLASS="FILENAME"
>contrib</TT
> to <TT
CLASS="FILENAME"
>src/test/modules</TT
>
       (&Aacute;lvaro Herrera)
      </P
><P
>       These modules are only meant for server testing, so they do not need
       to be built or installed when packaging <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>.
      </P
></LI
></UL
></DIV
></DIV
></DIV
><DIV
CLASS="NAVFOOTER"
><HR
ALIGN="LEFT"
WIDTH="100%"><TABLE
SUMMARY="Footer navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
><A
HREF="release-9-5-1.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="index.html"
ACCESSKEY="H"
>Home</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
><A
HREF="release-9-4-19.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>Release 9.5.1</TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="release.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
>Release 9.4.19</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>