path: root/doc/src/sgml/postgres-fdw.sgml

<!-- doc/src/sgml/postgres-fdw.sgml -->

<sect1 id="postgres-fdw" xreflabel="postgres_fdw">
 <title>postgres_fdw</title>

 <indexterm zone="postgres-fdw">
  <primary>postgres_fdw</primary>
 </indexterm>

 <para>
  The <filename>postgres_fdw</filename> module provides the foreign-data wrapper
  <literal>postgres_fdw</literal>, which can be used to access data
  stored in external <productname>PostgreSQL</productname> servers.
 </para>

 <para>
  The functionality provided by this module overlaps substantially
  with the functionality of the older <xref linkend="dblink"/> module.
  But <filename>postgres_fdw</filename> provides more transparent and
  standards-compliant syntax for accessing remote tables, and can give
  better performance in many cases.
 </para>

 <para>
  To prepare for remote access using <filename>postgres_fdw</filename>:
  <orderedlist spacing="compact">
   <listitem>
    <para>
     Install the  <filename>postgres_fdw</filename> extension using <xref
     linkend="sql-createextension"/>.
    </para>
   </listitem>
   <listitem>
    <para>
     Create a foreign server object, using <xref linkend="sql-createserver"/>,
     to represent each remote database you want to connect to.
     Specify connection information, except <literal>user</literal> and
     <literal>password</literal>, as options of the server object.
    </para>
   </listitem>
   <listitem>
    <para>
     Create a user mapping, using <xref linkend="sql-createusermapping"/>, for
     each database user you want to allow to access each foreign server.
     Specify the remote user name and password to use as
     <literal>user</literal> and <literal>password</literal> options of the
     user mapping.
    </para>
   </listitem>
   <listitem>
    <para>
     Create a foreign table, using <xref linkend="sql-createforeigntable"/>
     or <xref linkend="sql-importforeignschema"/>,
     for each remote table you want to access.  The columns of the foreign
     table must match the referenced remote table.  You can, however, use
     table and/or column names different from the remote table's, if you
     specify the correct remote names as options of the foreign table object.
    </para>
   </listitem>
  </orderedlist>
 </para>

 <para>
  Now you need only <command>SELECT</command> from a foreign table to access
  the data stored in its underlying remote table.  You can also modify
  the remote table using <command>INSERT</command>, <command>UPDATE</command>,
  <command>DELETE</command>, <command>COPY</command>, or
  <command>TRUNCATE</command>.
  (Of course, the remote user you have specified in your user mapping must
  have privileges to do these things.)
 </para>

 <para>
  Note that the <literal>ONLY</literal> option specified in
  <command>SELECT</command>, <command>UPDATE</command>,
  <command>DELETE</command> or <command>TRUNCATE</command>
  has no effect when accessing or modifying the remote table.
 </para>

 <para>
  Note that <filename>postgres_fdw</filename> currently lacks support for
  <command>INSERT</command> statements with an <literal>ON CONFLICT DO
  UPDATE</literal> clause.  However, the <literal>ON CONFLICT DO NOTHING</literal>
  clause is supported, provided a unique index inference specification
  is omitted.
  Note also that <filename>postgres_fdw</filename> supports row movement
  invoked by <command>UPDATE</command> statements executed on partitioned
  tables, but it currently does not handle the case where a remote partition
  chosen to insert a moved row into is also an <command>UPDATE</command>
  target partition that will be updated elsewhere in the same command.
 </para>

 <para>
  It is generally recommended that the columns of a foreign table be declared
  with exactly the same data types, and collations if applicable, as the
  referenced columns of the remote table.  Although <filename>postgres_fdw</filename>
  is currently rather forgiving about performing data type conversions at
  need, surprising semantic anomalies may arise when types or collations do
  not match, due to the remote server interpreting query conditions
  differently from the local server.
 </para>

 <para>
  Note that a foreign table can be declared with fewer columns, or with a
  different column order, than its underlying remote table has.  Matching
  of columns to the remote table is by name, not position.
 </para>

 <sect2>
  <title>FDW Options of postgres_fdw</title>

  <sect3>
   <title>Connection Options</title>

   <para>
    A foreign server using the <filename>postgres_fdw</filename> foreign data wrapper
    can have the same options that <application>libpq</application> accepts in
    connection strings, as described in <xref linkend="libpq-paramkeywords"/>,
    except that these options are not allowed or have special handling:

    <itemizedlist spacing="compact">
     <listitem>
      <para>
       <literal>user</literal>, <literal>password</literal> and <literal>sslpassword</literal> (specify these
       in a user mapping, instead, or use a service file)
      </para>
     </listitem>
     <listitem>
      <para>
       <literal>client_encoding</literal> (this is automatically set from the local
       server encoding)
      </para>
     </listitem>
     <listitem>
      <para>
       <literal>fallback_application_name</literal> (always set to
       <literal>postgres_fdw</literal>)
      </para>
     </listitem>
     <listitem>
      <para>
       <literal>sslkey</literal> and <literal>sslcert</literal> - these may
       appear in <emphasis>either or both</emphasis> a connection and a user
       mapping. If both are present, the user mapping setting overrides the
       connection setting.
      </para>
     </listitem>
    </itemizedlist>
   </para>

   <para>
    Only superusers may create or modify user mappings with the
    <literal>sslcert</literal> or <literal>sslkey</literal> settings.
   </para>
   <para>
    Only superusers may connect to foreign servers without password
    authentication, so always specify the <literal>password</literal> option
    for user mappings belonging to non-superusers.
   </para>
   <para>
    A superuser may override this check on a per-user-mapping basis by setting
    the user mapping option <literal>password_required 'false'</literal>, e.g.,
<programlisting>
ALTER USER MAPPING FOR some_non_superuser SERVER loopback_nopw
OPTIONS (ADD password_required 'false');
</programlisting>
    To prevent unprivileged users from exploiting the authentication rights
    of the unix user the postgres server is running as to escalate to superuser
    rights, only the superuser may set this option on a user mapping.
    </para>
    <para>
    Care is required to ensure that this does not allow the mapped
    user the ability to connect as superuser to the mapped database per
    CVE-2007-3278 and CVE-2007-6601. Don't set
    <literal>password_required=false</literal>
    on the <literal>public</literal> role. Keep in mind that the mapped
    user can potentially use any client certificates,
    <filename>.pgpass</filename>,
    <filename>.pg_service.conf</filename> etc in the unix home directory of the
    system user the postgres server runs as. They can also use any trust
    relationship granted by authentication modes like <literal>peer</literal>
    or <literal>ident</literal> authentication.
   </para>
  </sect3>

  <sect3>
   <title>Object Name Options</title>

   <para>
    These options can be used to control the names used in SQL statements
    sent to the remote <productname>PostgreSQL</productname> server.  These
    options are needed when a foreign table is created with names different
    from the underlying remote table's names.
   </para>

   <variablelist>

    <varlistentry>
     <term><literal>schema_name</literal></term>
     <listitem>
      <para>
       This option, which can be specified for a foreign table, gives the
       schema name to use for the foreign table on the remote server.  If this
       option is omitted, the name of the foreign table's schema is used.
      </para>
     </listitem>
    </varlistentry>

    <varlistentry>
     <term><literal>table_name</literal></term>
     <listitem>
      <para>
       This option, which can be specified for a foreign table, gives the
       table name to use for the foreign table on the remote server.  If this
       option is omitted, the foreign table's name is used.
      </para>
     </listitem>
    </varlistentry>

    <varlistentry>
     <term><literal>column_name</literal></term>
     <listitem>
      <para>
       This option, which can be specified for a column of a foreign table,
       gives the column name to use for the column on the remote server.
       If this option is omitted, the column's name is used.
      </para>
     </listitem>
    </varlistentry>

   </variablelist>

  </sect3>

  <sect3>
   <title>Cost Estimation Options</title>

   <para>
    <filename>postgres_fdw</filename> retrieves remote data by executing queries
    against remote servers, so ideally the estimated cost of scanning a
    foreign table should be whatever it costs to be done on the remote
    server, plus some overhead for communication.  The most reliable way to
    get such an estimate is to ask the remote server and then add something
    for overhead &mdash; but for simple queries, it may not be worth the cost
    of an additional remote query to get a cost estimate.
    So <filename>postgres_fdw</filename> provides the following options to control
    how cost estimation is done:
   </para>

   <variablelist>

    <varlistentry>
     <term><literal>use_remote_estimate</literal></term>
     <listitem>
      <para>
       This option, which can be specified for a foreign table or a foreign
       server, controls whether <filename>postgres_fdw</filename> issues remote
       <command>EXPLAIN</command> commands to obtain cost estimates.
       A setting for a foreign table overrides any setting for its server,
       but only for that table.
       The default is <literal>false</literal>.
      </para>
     </listitem>
    </varlistentry>

    <varlistentry>
     <term><literal>fdw_startup_cost</literal></term>
     <listitem>
      <para>
       This option, which can be specified for a foreign server, is a floating
       point value that is added to the estimated startup cost of any
       foreign-table scan on that server.  This represents the additional
       overhead of establishing a connection, parsing and planning the query on
       the remote side, etc.
       The default value is <literal>100</literal>.
      </para>
     </listitem>
    </varlistentry>

    <varlistentry>
     <term><literal>fdw_tuple_cost</literal></term>
     <listitem>
      <para>
       This option, which can be specified for a foreign server, is a floating
       point value that is used as extra cost per-tuple for foreign-table
       scans on that server.  This represents the additional overhead of
       data transfer between servers.  You might increase or decrease this
       number to reflect higher or lower network delay to the remote server.
       The default value is <literal>0.01</literal>.
      </para>
     </listitem>
    </varlistentry>

   </variablelist>

   <para>
    When <literal>use_remote_estimate</literal> is true,
    <filename>postgres_fdw</filename> obtains row count and cost estimates from the
    remote server and then adds <literal>fdw_startup_cost</literal> and
    <literal>fdw_tuple_cost</literal> to the cost estimates.  When
    <literal>use_remote_estimate</literal> is false,
    <filename>postgres_fdw</filename> performs local row count and cost estimation
    and then adds <literal>fdw_startup_cost</literal> and
    <literal>fdw_tuple_cost</literal> to the cost estimates.  This local
    estimation is unlikely to be very accurate unless local copies of the
    remote table's statistics are available.  Running
    <xref linkend="sql-analyze"/> on the foreign table is the way to update
    the local statistics; this will perform a scan of the remote table and
    then calculate and store statistics just as though the table were local.
    Keeping local statistics can be a useful way to reduce per-query planning
    overhead for a remote table &mdash; but if the remote table is
    frequently updated, the local statistics will soon be obsolete.
   </para>

  </sect3>

  <sect3>
   <title>Remote Execution Options</title>

   <para>
    By default, only <literal>WHERE</literal> clauses using built-in operators and
    functions will be considered for execution on the remote server.  Clauses
    involving non-built-in functions are checked locally after rows are
    fetched.  If such functions are available on the remote server and can be
    relied on to produce the same results as they do locally, performance can
    be improved by sending such <literal>WHERE</literal> clauses for remote
    execution.  This behavior can be controlled using the following option:
   </para>

   <variablelist>

    <varlistentry>
     <term><literal>extensions</literal></term>
     <listitem>
      <para>
       This option is a comma-separated list of names
       of <productname>PostgreSQL</productname> extensions that are installed, in
       compatible versions, on both the local and remote servers.  Functions
       and operators that are immutable and belong to a listed extension will
       be considered shippable to the remote server.
       This option can only be specified for foreign servers, not per-table.
      </para>

      <para>
       When using the <literal>extensions</literal> option, <emphasis>it is the
       user's responsibility</emphasis> that the listed extensions exist and behave
       identically on both the local and remote servers.  Otherwise, remote
       queries may fail or behave unexpectedly.
      </para>
     </listitem>
    </varlistentry>

    <varlistentry>
     <term><literal>fetch_size</literal></term>
     <listitem>
      <para>
       This option specifies the number of rows <filename>postgres_fdw</filename>
       should get in each fetch operation. It can be specified for a foreign
       table or a foreign server. The option specified on a table overrides
       an option specified for the server.
       The default is <literal>100</literal>.
      </para>
     </listitem>
    </varlistentry>

    <varlistentry>
     <term><literal>batch_size</literal></term>
     <listitem>
      <para>
       This option specifies the number of rows <filename>postgres_fdw</filename>
       should insert in each insert operation. It can be specified for a
       foreign table or a foreign server. The option specified on a table
       overrides an option specified for the server.
       The default is <literal>1</literal>.
      </para>

      <para>
       Note the actual number of rows <filename>postgres_fdw</filename> inserts at
       once depends on the number of columns and the provided
       <literal>batch_size</literal> value. The batch is executed as a single
       query, and the libpq protocol (which <filename>postgres_fdw</filename>
       uses to connect to a remote server) limits the number of parameters in a
       single query to 65535. When the number of columns * <literal>batch_size</literal>
       exceeds the limit, the <literal>batch_size</literal> will be adjusted to
       avoid an error.
      </para>
     </listitem>
    </varlistentry>

   </variablelist>

  </sect3>

  <sect3>
   <title>Asynchronous Execution Options</title>

   <para>
    <filename>postgres_fdw</filename> supports asynchronous execution, which
    runs multiple parts of an <structname>Append</structname> node
    concurrently rather than serially to improve performance.
    This execution can be controlled using the following option:
   </para>

   <variablelist>

    <varlistentry>
     <term><literal>async_capable</literal></term>
     <listitem>
      <para>
       This option controls whether <filename>postgres_fdw</filename> allows
       foreign tables to be scanned concurrently for asynchronous execution.
       It can be specified for a foreign table or a foreign server.
       A table-level option overrides a server-level option.
       The default is <literal>false</literal>.
      </para>

      <para>
       In order to ensure that the data being returned from a foreign server
       is consistent, <filename>postgres_fdw</filename> will only open one
       connection for a given foreign server and will run all queries against
       that server sequentially even if there are multiple foreign tables
       involved, unless those tables are subject to different user mappings.
       In such a case, it may be more performant to disable this option to
       eliminate the overhead associated with running queries asynchronously.
      </para>

      <para>
       Asynchronous execution is applied even when an
       <structname>Append</structname> node contains subplan(s) executed
       synchronously as well as subplan(s) executed asynchronously.
       In such a case, if the asynchronous subplans are ones processed using
       <filename>postgres_fdw</filename>, tuples from the asynchronous
       subplans are not returned until after at least one synchronous subplan
       returns all tuples, as that subplan is executed while the asynchronous
       subplans are waiting for the results of asynchronous queries sent to
       foreign servers.
       This behavior might change in a future release.
      </para>
     </listitem>
    </varlistentry>

   </variablelist>
  </sect3>

  <sect3>
   <title>Updatability Options</title>

   <para>
    By default all foreign tables using <filename>postgres_fdw</filename> are assumed
    to be updatable.  This may be overridden using the following option:
   </para>

   <variablelist>

    <varlistentry>
     <term><literal>updatable</literal></term>
     <listitem>
      <para>
       This option controls whether <filename>postgres_fdw</filename> allows foreign
       tables to be modified using <command>INSERT</command>, <command>UPDATE</command> and
       <command>DELETE</command> commands.  It can be specified for a foreign table
       or a foreign server.  A table-level option overrides a server-level
       option.
       The default is <literal>true</literal>.
      </para>

      <para>
       Of course, if the remote table is not in fact updatable, an error
       would occur anyway.  Use of this option primarily allows the error to
       be thrown locally without querying the remote server.  Note however
       that the <literal>information_schema</literal> views will report a
       <filename>postgres_fdw</filename> foreign table to be updatable (or not)
       according to the setting of this option, without any check of the
       remote server.
      </para>
     </listitem>
    </varlistentry>

   </variablelist>
  </sect3>

  <sect3>
   <title>Truncatability Options</title>

   <para>
    By default all foreign tables using <filename>postgres_fdw</filename> are assumed
    to be truncatable.  This may be overridden using the following option:
   </para>

   <variablelist>

    <varlistentry>
     <term><literal>truncatable</literal></term>
     <listitem>
      <para>
       This option controls whether <filename>postgres_fdw</filename> allows
       foreign tables to be truncated using the <command>TRUNCATE</command>
       command. It can be specified for a foreign table or a foreign server.
       A table-level option overrides a server-level option.
       The default is <literal>true</literal>.
      </para>

      <para>
       Of course, if the remote table is not in fact truncatable, an error
       would occur anyway.  Use of this option primarily allows the error to
       be thrown locally without querying the remote server.
      </para>
     </listitem>
    </varlistentry>
   </variablelist>
  </sect3>

  <sect3>
   <title>Importing Options</title>

   <para>
    <filename>postgres_fdw</filename> is able to import foreign table definitions
    using <xref linkend="sql-importforeignschema"/>.  This command creates
    foreign table definitions on the local server that match tables or
    views present on the remote server.  If the remote tables to be imported
    have columns of user-defined data types, the local server must have
    compatible types of the same names.
   </para>

   <para>
    Importing behavior can be customized with the following options
    (given in the <command>IMPORT FOREIGN SCHEMA</command> command):
   </para>

   <variablelist>
    <varlistentry>
     <term><literal>import_collate</literal></term>
     <listitem>
      <para>
       This option controls whether column <literal>COLLATE</literal> options
       are included in the definitions of foreign tables imported
       from a foreign server. The default is <literal>true</literal>.  You might
       need to turn this off if the remote server has a different set of
       collation names than the local server does, which is likely to be the
       case if it's running on a different operating system.
       If you do so, however, there is a very severe risk that the imported
       table columns' collations will not match the underlying data, resulting
       in anomalous query behavior.
      </para>

      <para>
       Even when this parameter is set to <literal>true</literal>, importing
       columns whose collation is the remote server's default can be risky.
       They will be imported with <literal>COLLATE "default"</literal>, which
       will select the local server's default collation, which could be
       different.
      </para>
     </listitem>
    </varlistentry>
    <varlistentry>
     <term><literal>import_default</literal></term>
     <listitem>
      <para>
       This option controls whether column <literal>DEFAULT</literal> expressions
       are included in the definitions of foreign tables imported
       from a foreign server. The default is <literal>false</literal>.  If you
       enable this option, be wary of defaults that might get computed
       differently on the local server than they would be on the remote
       server; <function>nextval()</function> is a common source of problems.
       The <command>IMPORT</command> will fail altogether if an imported default
       expression uses a function or operator that does not exist locally.
      </para>
     </listitem>
    </varlistentry>
    <varlistentry>
     <term><literal>import_generated</literal></term>
     <listitem>
      <para>
       This option controls whether column <literal>GENERATED</literal> expressions
       are included in the definitions of foreign tables imported
       from a foreign server. The default is <literal>true</literal>.
       The <command>IMPORT</command> will fail altogether if an imported generated
       expression uses a function or operator that does not exist locally.
      </para>
     </listitem>
    </varlistentry>
    <varlistentry>
     <term><literal>import_not_null</literal></term>
     <listitem>
      <para>
       This option controls whether column <literal>NOT NULL</literal>
       constraints are included in the definitions of foreign tables imported
       from a foreign server. The default is <literal>true</literal>.
      </para>
     </listitem>
    </varlistentry>
   </variablelist>

   <para>
    Note that constraints other than <literal>NOT NULL</literal> will never be
    imported from the remote tables.  Although <productname>PostgreSQL</productname>
    does support check constraints on foreign tables, there is no
    provision for importing them automatically, because of the risk that a
    constraint expression could evaluate differently on the local and remote
    servers.  Any such inconsistency in the behavior of a check
    constraint could lead to hard-to-detect errors in query optimization.
    So if you wish to import check constraints, you must do so
    manually, and you should verify the semantics of each one carefully.
    For more detail about the treatment of check constraints on
    foreign tables, see <xref linkend="sql-createforeigntable"/>.
   </para>

   <para>
    Tables or foreign tables which are partitions of some other table are
    imported only when they are explicitly specified in
    <literal>LIMIT TO</literal> clause.  Otherwise they are automatically
    excluded from <xref linkend="sql-importforeignschema"/>.
    Since all data can be accessed through the partitioned table
    which is the root of the partitioning hierarchy, importing only
    partitioned tables should allow access to all the data without
    creating extra objects.
   </para>

  </sect3>

  <sect3>
    <title>Connection Management Options</title>

    <para>
     By default, all connections that <filename>postgres_fdw</filename>
     establishes to foreign servers are kept open in the local session
     for re-use.
    </para>

    <variablelist>

     <varlistentry>
      <term><literal>keep_connections</literal></term>
      <listitem>
       <para>
        This option controls whether <filename>postgres_fdw</filename> keeps
        the connections to the foreign server open so that subsequent
        queries can re-use them. It can only be specified for a foreign server.
        The default is <literal>on</literal>. If set to <literal>off</literal>,
        all connections to this foreign server will be discarded at the end of
        each transaction.
      </para>
      </listitem>
     </varlistentry>

    </variablelist>
   </sect3>
 </sect2>

<sect2>
  <title>Functions</title>

  <variablelist>
   <varlistentry>
    <term><function>postgres_fdw_get_connections(OUT server_name text, OUT valid boolean) returns setof record</function></term>
    <listitem>
     <para>
      This function returns the foreign server names of all the open
      connections that <filename>postgres_fdw</filename> established from
      the local session to the foreign servers. It also returns whether
      each connection is valid or not. <literal>false</literal> is returned
      if the foreign server connection is used in the current local
      transaction but its foreign server or user mapping is changed or
      dropped (Note that server name of an invalid connection will be
      <literal>NULL</literal> if the server is dropped),
      and then such invalid connection will be closed at
      the end of that transaction. <literal>true</literal> is returned
      otherwise. If there are no open connections, no record is returned.
      Example usage of the function:
<screen>
postgres=# SELECT * FROM postgres_fdw_get_connections() ORDER BY 1;
 server_name | valid 
-------------+-------
 loopback1   | t
 loopback2   | f
</screen>
     </para>
    </listitem>
   </varlistentry>

   <varlistentry>
    <term><function>postgres_fdw_disconnect(server_name text) returns boolean</function></term>
    <listitem>
     <para>
      This function discards the open connections that are established by
      <filename>postgres_fdw</filename> from the local session to
      the foreign server with the given name.  Note that there can be
      multiple connections to the given server using different user mappings.
      If the connections are used in the current local transaction,
      they are not disconnected and warning messages are reported.
      This function returns <literal>true</literal> if it disconnects
      at least one connection, otherwise <literal>false</literal>.
      If no foreign server with the given name is found, an error is reported.
      Example usage of the function:
<screen>
postgres=# SELECT postgres_fdw_disconnect('loopback1');
 postgres_fdw_disconnect 
-------------------------
 t
</screen>
     </para>
    </listitem>
   </varlistentry>

   <varlistentry>
    <term><function>postgres_fdw_disconnect_all() returns boolean</function></term>
    <listitem>
     <para>
      This function discards all the open connections that are established by
      <filename>postgres_fdw</filename> from the local session to
      foreign servers.  If the connections are used in the current local
      transaction, they are not disconnected and warning messages are reported.
      This function returns <literal>true</literal> if it disconnects
      at least one connection, otherwise <literal>false</literal>.
      Example usage of the function:
<screen>
postgres=# SELECT postgres_fdw_disconnect_all();
 postgres_fdw_disconnect_all 
-----------------------------
 t
</screen>
     </para>
    </listitem>
   </varlistentry>
   </variablelist>

</sect2>

 <sect2>
  <title>Connection Management</title>

  <para>
   <filename>postgres_fdw</filename> establishes a connection to a
   foreign server during the first query that uses a foreign table
   associated with the foreign server.  By default this connection
   is kept and re-used for subsequent queries in the same session.
   This behavior can be controlled using
   <literal>keep_connections</literal> option for a foreign server. If
   multiple user identities (user mappings) are used to access the foreign
   server, a connection is established for each user mapping.
  </para>

  <para>
   When changing the definition of or removing a foreign server or
   a user mapping, the associated connections are closed.
   But note that if any connections are in use in the current local transaction,
   they are kept until the end of the transaction.
   Closed connections will be re-established when they are necessary
   by future queries using a foreign table.
  </para>

  <para>
   Once a connection to a foreign server has been established,
   it's by default kept until the local or corresponding remote
   session exits.  To disconnect a connection explicitly,
   <literal>keep_connections</literal> option for a foreign server
   may be disabled, or
   <function>postgres_fdw_disconnect</function> and
   <function>postgres_fdw_disconnect_all</function> functions
   may be used.  For example, these are useful to close
   connections that are no longer necessary, thereby releasing
   connections on the foreign server.
  </para>
 </sect2>

 <sect2>
  <title>Transaction Management</title>

  <para>
   During a query that references any remote tables on a foreign server,
   <filename>postgres_fdw</filename> opens a transaction on the
   remote server if one is not already open corresponding to the current
   local transaction.  The remote transaction is committed or aborted when
   the local transaction commits or aborts.  Savepoints are similarly
   managed by creating corresponding remote savepoints.
  </para>

  <para>
   The remote transaction uses <literal>SERIALIZABLE</literal>
   isolation level when the local transaction has <literal>SERIALIZABLE</literal>
   isolation level; otherwise it uses <literal>REPEATABLE READ</literal>
   isolation level.  This choice ensures that if a query performs multiple
   table scans on the remote server, it will get snapshot-consistent results
   for all the scans.  A consequence is that successive queries within a
   single transaction will see the same data from the remote server, even if
   concurrent updates are occurring on the remote server due to other
   activities.  That behavior would be expected anyway if the local
   transaction uses <literal>SERIALIZABLE</literal> or <literal>REPEATABLE READ</literal>
   isolation level, but it might be surprising for a <literal>READ
   COMMITTED</literal> local transaction.  A future
   <productname>PostgreSQL</productname> release might modify these rules.
  </para>

  <para>
   Note that it is currently not supported by
   <filename>postgres_fdw</filename> to prepare the remote transaction for
   two-phase commit.
  </para>
 </sect2>

 <sect2>
  <title>Remote Query Optimization</title>

  <para>
   <filename>postgres_fdw</filename> attempts to optimize remote queries to reduce
   the amount of data transferred from foreign servers.  This is done by
   sending query <literal>WHERE</literal> clauses to the remote server for
   execution, and by not retrieving table columns that are not needed for
   the current query.  To reduce the risk of misexecution of queries,
   <literal>WHERE</literal> clauses are not sent to the remote server unless they use
   only data types, operators, and functions that are built-in or belong to an
   extension that's listed in the foreign server's <literal>extensions</literal>
   option.  Operators and functions in such clauses must
   be <literal>IMMUTABLE</literal> as well.
   For an <command>UPDATE</command> or <command>DELETE</command> query,
   <filename>postgres_fdw</filename> attempts to optimize the query execution by
   sending the whole query to the remote server if there are no query
   <literal>WHERE</literal> clauses that cannot be sent to the remote server,
   no local joins for the query, no row-level local <literal>BEFORE</literal> or
   <literal>AFTER</literal> triggers or stored generated columns on the target
   table, and no <literal>CHECK OPTION</literal> constraints from parent
   views.  In <command>UPDATE</command>,
   expressions to assign to target columns must use only built-in data types,
   <literal>IMMUTABLE</literal> operators, or <literal>IMMUTABLE</literal> functions,
   to reduce the risk of misexecution of the query.
  </para>

  <para>
   When <filename>postgres_fdw</filename> encounters a join between foreign tables on
   the same foreign server, it sends the entire join to the foreign server,
   unless for some reason it believes that it will be more efficient to fetch
   rows from each table individually, or unless the table references involved
   are subject to different user mappings.  While sending the <literal>JOIN</literal>
   clauses, it takes the same precautions as mentioned above for the
   <literal>WHERE</literal> clauses.
  </para>

  <para>
   The query that is actually sent to the remote server for execution can
   be examined using <command>EXPLAIN VERBOSE</command>.
  </para>
 </sect2>

 <sect2>
  <title>Remote Query Execution Environment</title>

  <para>
   In the remote sessions opened by <filename>postgres_fdw</filename>,
   the <xref linkend="guc-search-path"/> parameter is set to
   just <literal>pg_catalog</literal>, so that only built-in objects are visible
   without schema qualification.  This is not an issue for queries
   generated by <filename>postgres_fdw</filename> itself, because it always
   supplies such qualification.  However, this can pose a hazard for
   functions that are executed on the remote server via triggers or rules
   on remote tables.  For example, if a remote table is actually a view,
   any functions used in that view will be executed with the restricted
   search path.  It is recommended to schema-qualify all names in such
   functions, or else attach <literal>SET search_path</literal> options
   (see <xref linkend="sql-createfunction"/>) to such functions
   to establish their expected search path environment.
  </para>

  <para>
   <filename>postgres_fdw</filename> likewise establishes remote session settings
   for various parameters:
   <itemizedlist spacing="compact">
    <listitem>
     <para>
      <xref linkend="guc-timezone"/> is set to <literal>UTC</literal>
     </para>
    </listitem>
    <listitem>
     <para>
      <xref linkend="guc-datestyle"/> is set to <literal>ISO</literal>
     </para>
    </listitem>
    <listitem>
     <para>
      <xref linkend="guc-intervalstyle"/> is set to <literal>postgres</literal>
     </para>
    </listitem>
    <listitem>
     <para>
      <xref linkend="guc-extra-float-digits"/> is set to <literal>3</literal> for remote
      servers 9.0 and newer and is set to <literal>2</literal> for older versions
     </para>
    </listitem>
   </itemizedlist>
   These are less likely to be problematic than <varname>search_path</varname>, but
   can be handled with function <literal>SET</literal> options if the need arises.
  </para>

  <para>
   It is <emphasis>not</emphasis> recommended that you override this behavior by
   changing the session-level settings of these parameters; that is likely
   to cause <filename>postgres_fdw</filename> to malfunction.
  </para>
 </sect2>

 <sect2>
  <title>Cross-Version Compatibility</title>

  <para>
   <filename>postgres_fdw</filename> can be used with remote servers dating back
   to <productname>PostgreSQL</productname> 8.3.  Read-only capability is available
   back to 8.1.  A limitation however is that <filename>postgres_fdw</filename>
   generally assumes that immutable built-in functions and operators are
   safe to send to the remote server for execution, if they appear in a
   <literal>WHERE</literal> clause for a foreign table.  Thus, a built-in
   function that was added since the remote server's release might be sent
   to it for execution, resulting in <quote>function does not exist</quote> or
   a similar error.  This type of failure can be worked around by
   rewriting the query, for example by embedding the foreign table
   reference in a sub-<literal>SELECT</literal> with <literal>OFFSET 0</literal> as an
   optimization fence, and placing the problematic function or operator
   outside the sub-<literal>SELECT</literal>.
  </para>
 </sect2>

 <sect2>
  <title>Examples</title>

  <para>
   Here is an example of creating a foreign table with
   <literal>postgres_fdw</literal>. First install the extension:
  </para>

<programlisting>
CREATE EXTENSION postgres_fdw;
</programlisting>

  <para>
   Then create a foreign server using <xref linkend="sql-createserver"/>.
   In this example we wish to connect to a <productname>PostgreSQL</productname> server
   on host <literal>192.83.123.89</literal> listening on
   port <literal>5432</literal>.  The database to which the connection is made
   is named <literal>foreign_db</literal> on the remote server:

<programlisting>
CREATE SERVER foreign_server
        FOREIGN DATA WRAPPER postgres_fdw
        OPTIONS (host '192.83.123.89', port '5432', dbname 'foreign_db');
</programlisting>
  </para>

  <para>
   A user mapping, defined with <xref linkend="sql-createusermapping"/>, is
   needed as well to identify the role that will be used on the remote
   server:

<programlisting>
CREATE USER MAPPING FOR local_user
        SERVER foreign_server
        OPTIONS (user 'foreign_user', password 'password');
</programlisting>
  </para>

  <para>
   Now it is possible to create a foreign table with
   <xref linkend="sql-createforeigntable"/>.  In this example we
   wish to access the table named <structname>some_schema.some_table</structname>
   on the remote server.  The local name for it will
   be <structname>foreign_table</structname>:

<programlisting>
CREATE FOREIGN TABLE foreign_table (
        id integer NOT NULL,
        data text
)
        SERVER foreign_server
        OPTIONS (schema_name 'some_schema', table_name 'some_table');
</programlisting>

   It's essential that the data types and other properties of the columns
   declared in <command>CREATE FOREIGN TABLE</command> match the actual remote table.
   Column names must match as well, unless you attach <literal>column_name</literal>
   options to the individual columns to show how they are named in the remote
   table.
   In many cases, use of <link linkend="sql-importforeignschema"><command>IMPORT FOREIGN SCHEMA</command></link> is
   preferable to constructing foreign table definitions manually.
  </para>
 </sect2>

 <sect2>
  <title>Author</title>
  <para>
   Shigeru Hanada <email>shigeru.hanada@gmail.com</email>
  </para>
 </sect2>

</sect1>