Baoding-FenShuaJiang
/
Green_GoodERP18_FSJ_Standard
Suivre 1
Ajouter aux favoris 0
Wiki 捐赠
gooderp18绿色标准版
Vous ne pouvez pas sélectionner plus de 25 sujets Les noms de sujets doivent commencer par une lettre ou un nombre, peuvent contenir des tirets ('-') et peuvent comporter jusqu'à 35 caractères.
Aborescence: 0f75e31513
Branches Tags
${ item.name }
Créer la branche ${ searchTerm }
de '0f75e31513'
${ noResults }
Green_GoodERP18_FSJ_Standard/PostgreSQL/doc/postgresql/html/app-pgdump.html

786 lignes
60KB
Brut Annotations Historique 

  1. <?xml version="1.0" encoding="UTF-8" standalone="no"?>

  2. <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>pg_dump</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets V1.79.1" /><link rel="prev" href="app-pgconfig.html" title="pg_config" /><link rel="next" href="app-pg-dumpall.html" title="pg_dumpall" /></head><body><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">pg_dump</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="app-pgconfig.html" title="pg_config">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="reference-client.html" title="PostgreSQL Client Applications">Up</a></td><th width="60%" align="center">PostgreSQL Client Applications</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 12.4 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="app-pg-dumpall.html" title="pg_dumpall">Next</a></td></tr></table><hr></hr></div><div class="refentry" id="APP-PGDUMP"><div class="titlepage"></div><a id="id-1.9.4.12.1" class="indexterm"></a><div class="refnamediv"><h2><span class="refentrytitle">pg_dump</span></h2><p>pg_dump — 

  3.    extract a <span class="productname">PostgreSQL</span> database into a script file or other archive file

  4.   </p></div><div class="refsynopsisdiv"><h2>Synopsis</h2><div class="cmdsynopsis"><p id="id-1.9.4.12.4.1"><code class="command">pg_dump</code> [<em class="replaceable"><code>connection-option</code></em>...] [<em class="replaceable"><code>option</code></em>...] [<em class="replaceable"><code>dbname</code></em>]</p></div></div><div class="refsect1" id="PG-DUMP-DESCRIPTION"><h2>

  5.    Description

  6.   </h2><p>

  7.    <span class="application">pg_dump</span> is a utility for backing up a

  8.    <span class="productname">PostgreSQL</span> database. It makes consistent

  9.    backups even if the database is being used concurrently.

  10.    <span class="application">pg_dump</span> does not block other users

  11.    accessing the database (readers or writers).

  12.   </p><p>

  13.    <span class="application">pg_dump</span> only dumps a single database.

  14.    To back up an entire cluster, or to back up global objects that are

  15.    common to all databases in a cluster (such as roles and tablespaces),

  16.    use <a class="xref" href="app-pg-dumpall.html" title="pg_dumpall"><span class="refentrytitle"><span class="application">pg_dumpall</span></span></a>.

  17.   </p><p>

  18.    Dumps can be output in script or archive file formats. Script

  19.    dumps are plain-text files containing the SQL commands required

  20.    to reconstruct the database to the state it was in at the time it was

  21.    saved. To restore from such a script, feed it to <a class="xref" href="app-psql.html" title="psql"><span class="refentrytitle"><span class="application">psql</span></span></a>. Script files

  22.    can be used to reconstruct the database even on other machines and

  23.    other architectures; with some modifications, even on other SQL

  24.    database products.

  25.   </p><p>

  26.    The alternative archive file formats must be used with

  27.    <a class="xref" href="app-pgrestore.html" title="pg_restore"><span class="refentrytitle">pg_restore</span></a> to rebuild the database.  They

  28.    allow <span class="application">pg_restore</span> to be selective about

  29.    what is restored, or even to reorder the items prior to being

  30.    restored.

  31.    The archive file formats are designed to be portable across

  32.    architectures.

  33.   </p><p>

  34.    When used with one of the archive file formats and combined with

  35.    <span class="application">pg_restore</span>,

  36.    <span class="application">pg_dump</span> provides a flexible archival and

  37.    transfer mechanism. <span class="application">pg_dump</span> can be used to

  38.    backup an entire database, then <span class="application">pg_restore</span>

  39.    can be used to examine the archive and/or select which parts of the

  40.    database are to be restored. The most flexible output file formats are

  41.    the <span class="quote">“<span class="quote">custom</span>”</span> format (<code class="option">-Fc</code>) and the

  42.    <span class="quote">“<span class="quote">directory</span>”</span> format (<code class="option">-Fd</code>). They allow

  43.    for selection and reordering of all archived items, support parallel

  44.    restoration, and are compressed by default. The <span class="quote">“<span class="quote">directory</span>”</span>

  45.    format is the only format that supports parallel dumps.

  46.   </p><p>

  47.    While running <span class="application">pg_dump</span>, one should examine the

  48.    output for any warnings (printed on standard error), especially in

  49.    light of the limitations listed below.

  50.   </p></div><div class="refsect1" id="PG-DUMP-OPTIONS"><h2>Options</h2><p>

  51.     The following command-line options control the content and

  52.     format of the output.

  53. 

  54.     </p><div class="variablelist"><dl class="variablelist"><dt><span class="term"><em class="replaceable"><code>dbname</code></em></span></dt><dd><p>

  55.         Specifies the name of the database to be dumped.  If this is

  56.         not specified, the environment variable

  57.         <code class="envar">PGDATABASE</code> is used.  If that is not set, the

  58.         user name specified for the connection is used.

  59.        </p></dd><dt><span class="term"><code class="option">-a</code><br /></span><span class="term"><code class="option">--data-only</code></span></dt><dd><p>

  60.         Dump only the data, not the schema (data definitions).

  61.         Table data, large objects, and sequence values are dumped.

  62.        </p><p>

  63.         This option is similar to, but for historical reasons not identical

  64.         to, specifying <code class="option">--section=data</code>.

  65.        </p></dd><dt><span class="term"><code class="option">-b</code><br /></span><span class="term"><code class="option">--blobs</code></span></dt><dd><p>

  66.         Include large objects in the dump.  This is the default behavior

  67.         except when <code class="option">--schema</code>, <code class="option">--table</code>, or

  68.         <code class="option">--schema-only</code> is specified.  The <code class="option">-b</code>

  69.         switch is therefore only useful to add large objects to dumps

  70.         where a specific schema or table has been requested.  Note that

  71.         blobs are considered data and therefore will be included when

  72.         <code class="option">--data-only</code> is used, but not

  73.         when <code class="option">--schema-only</code> is.

  74.        </p></dd><dt><span class="term"><code class="option">-B</code><br /></span><span class="term"><code class="option">--no-blobs</code></span></dt><dd><p>

  75.         Exclude large objects in the dump.

  76.        </p><p>

  77.         When both <code class="option">-b</code> and <code class="option">-B</code> are given, the behavior

  78.         is to output large objects, when data is being dumped, see the

  79.         <code class="option">-b</code> documentation.

  80.        </p></dd><dt><span class="term"><code class="option">-c</code><br /></span><span class="term"><code class="option">--clean</code></span></dt><dd><p>

  81.         Output commands to clean (drop)

  82.         database objects prior to outputting the commands for creating them.

  83.         (Unless <code class="option">--if-exists</code> is also specified,

  84.         restore might generate some harmless error messages, if any objects

  85.         were not present in the destination database.)

  86.        </p><p>

  87.         This option is only meaningful for the plain-text format.  For

  88.         the archive formats, you can specify the option when you

  89.         call <code class="command">pg_restore</code>.

  90.        </p></dd><dt><span class="term"><code class="option">-C</code><br /></span><span class="term"><code class="option">--create</code></span></dt><dd><p>

  91.         Begin the output with a command to create the

  92.         database itself and reconnect to the created database.  (With a

  93.         script of this form, it doesn't matter which database in the

  94.         destination installation you connect to before running the script.)

  95.         If <code class="option">--clean</code> is also specified, the script drops and

  96.         recreates the target database before reconnecting to it.

  97.        </p><p>

  98.         With <code class="option">--create</code>, the output also includes the

  99.         database's comment if any, and any configuration variable settings

  100.         that are specific to this database, that is,

  101.         any <code class="command">ALTER DATABASE ... SET ...</code>

  102.         and <code class="command">ALTER ROLE ... IN DATABASE ... SET ...</code>

  103.         commands that mention this database.

  104.         Access privileges for the database itself are also dumped,

  105.         unless <code class="option">--no-acl</code> is specified.

  106.        </p><p>

  107.         This option is only meaningful for the plain-text format.  For

  108.         the archive formats, you can specify the option when you

  109.         call <code class="command">pg_restore</code>.

  110.        </p></dd><dt><span class="term"><code class="option">-E <em class="replaceable"><code>encoding</code></em></code><br /></span><span class="term"><code class="option">--encoding=<em class="replaceable"><code>encoding</code></em></code></span></dt><dd><p>

  111.         Create the dump in the specified character set encoding. By default,

  112.         the dump is created in the database encoding.  (Another way to get the

  113.         same result is to set the <code class="envar">PGCLIENTENCODING</code> environment

  114.         variable to the desired dump encoding.)

  115.        </p></dd><dt><span class="term"><code class="option">-f <em class="replaceable"><code>file</code></em></code><br /></span><span class="term"><code class="option">--file=<em class="replaceable"><code>file</code></em></code></span></dt><dd><p>

  116.         Send output to the specified file. This parameter can be omitted for

  117.         file based output formats, in which case the standard output is used.

  118.         It must be given for the directory output format however, where it

  119.         specifies the target directory instead of a file. In this case the

  120.         directory is created by <code class="command">pg_dump</code> and must not exist

  121.         before.

  122.        </p></dd><dt><span class="term"><code class="option">-F <em class="replaceable"><code>format</code></em></code><br /></span><span class="term"><code class="option">--format=<em class="replaceable"><code>format</code></em></code></span></dt><dd><p>

  123.         Selects the format of the output.

  124.         <em class="replaceable"><code>format</code></em> can be one of the following:

  125. 

  126.        </p><div class="variablelist"><dl class="variablelist"><dt><span class="term"><code class="literal">p</code><br /></span><span class="term"><code class="literal">plain</code></span></dt><dd><p>

  127.            Output a plain-text <acronym class="acronym">SQL</acronym> script file (the default).

  128.           </p></dd><dt><span class="term"><code class="literal">c</code><br /></span><span class="term"><code class="literal">custom</code></span></dt><dd><p>

  129.            Output a custom-format archive suitable for input into

  130.            <span class="application">pg_restore</span>.

  131.            Together with the directory output format, this is the most flexible

  132.            output format in that it allows manual selection and reordering of

  133.            archived items during restore. This format is also compressed by

  134.            default.

  135.           </p></dd><dt><span class="term"><code class="literal">d</code><br /></span><span class="term"><code class="literal">directory</code></span></dt><dd><p>

  136.            Output a directory-format archive suitable for input into

  137.            <span class="application">pg_restore</span>. This will create a directory

  138.            with one file for each table and blob being dumped, plus a

  139.            so-called Table of Contents file describing the dumped objects in a

  140.            machine-readable format that <span class="application">pg_restore</span>

  141.            can read. A directory format archive can be manipulated with

  142.            standard Unix tools; for example, files in an uncompressed archive

  143.            can be compressed with the <span class="application">gzip</span> tool.

  144.            This format is compressed by default and also supports parallel

  145.            dumps.

  146.           </p></dd><dt><span class="term"><code class="literal">t</code><br /></span><span class="term"><code class="literal">tar</code></span></dt><dd><p>

  147.            Output a <code class="command">tar</code>-format archive suitable for input

  148.            into <span class="application">pg_restore</span>. The tar format is

  149.            compatible with the directory format: extracting a tar-format

  150.            archive produces a valid directory-format archive.

  151.            However, the tar format does not support compression. Also, when

  152.            using tar format the relative order of table data items cannot be

  153.            changed during restore.

  154.           </p></dd></dl></div></dd><dt><span class="term"><code class="option">-j <em class="replaceable"><code>njobs</code></em></code><br /></span><span class="term"><code class="option">--jobs=<em class="replaceable"><code>njobs</code></em></code></span></dt><dd><p>

  155.         Run the dump in parallel by dumping <em class="replaceable"><code>njobs</code></em>

  156.         tables simultaneously. This option reduces the time of the dump but it also

  157.         increases the load on the database server. You can only use this option with the

  158.         directory output format because this is the only output format where multiple processes

  159.         can write their data at the same time.

  160.        </p><p><span class="application">pg_dump</span> will open <em class="replaceable"><code>njobs</code></em>

  161.         + 1 connections to the database, so make sure your <a class="xref" href="runtime-config-connection.html#GUC-MAX-CONNECTIONS">max_connections</a>

  162.         setting is high enough to accommodate all connections.

  163.        </p><p>

  164.         Requesting exclusive locks on database objects while running a parallel dump could

  165.         cause the dump to fail. The reason is that the <span class="application">pg_dump</span> master process

  166.         requests shared locks on the objects that the worker processes are going to dump later

  167.         in order to

  168.         make sure that nobody deletes them and makes them go away while the dump is running.

  169.         If another client then requests an exclusive lock on a table, that lock will not be

  170.         granted but will be queued waiting for the shared lock of the master process to be

  171.         released. Consequently any other access to the table will not be granted either and

  172.         will queue after the exclusive lock request. This includes the worker process trying

  173.         to dump the table. Without any precautions this would be a classic deadlock situation.

  174.         To detect this conflict, the <span class="application">pg_dump</span> worker process requests another

  175.         shared lock using the <code class="literal">NOWAIT</code> option. If the worker process is not granted

  176.         this shared lock, somebody else must have requested an exclusive lock in the meantime

  177.         and there is no way to continue with the dump, so <span class="application">pg_dump</span> has no choice

  178.         but to abort the dump.

  179.        </p><p>

  180.         For a consistent backup, the database server needs to support

  181.         synchronized snapshots, a feature that was introduced in

  182.         <span class="productname">PostgreSQL</span> 9.2 for primary servers and 10

  183.         for standbys. With this feature, database clients can ensure they see

  184.         the same data set even though they use different connections.

  185.         <code class="command">pg_dump -j</code> uses multiple database connections; it

  186.         connects to the database once with the master process and once again

  187.         for each worker job. Without the synchronized snapshot feature, the

  188.         different worker jobs wouldn't be guaranteed to see the same data in

  189.         each connection, which could lead to an inconsistent backup.

  190.        </p><p>

  191.         If you want to run a parallel dump of a pre-9.2 server, you need to make sure that the

  192.         database content doesn't change from between the time the master connects to the

  193.         database until the last worker job has connected to the database. The easiest way to

  194.         do this is to halt any data modifying processes (DDL and DML) accessing the database

  195.         before starting the backup. You also need to specify the

  196.         <code class="option">--no-synchronized-snapshots</code> parameter when running

  197.         <code class="command">pg_dump -j</code> against a pre-9.2 <span class="productname">PostgreSQL</span>

  198.         server.

  199.        </p></dd><dt><span class="term"><code class="option">-n <em class="replaceable"><code>pattern</code></em></code><br /></span><span class="term"><code class="option">--schema=<em class="replaceable"><code>pattern</code></em></code></span></dt><dd><p>

  200.         Dump only schemas matching <em class="replaceable"><code>pattern</code></em>; this selects both the

  201.         schema itself, and all its contained objects.  When this option is

  202.         not specified, all non-system schemas in the target database will be

  203.         dumped.  Multiple schemas can be

  204.         selected by writing multiple <code class="option">-n</code> switches.  The

  205.         <em class="replaceable"><code>pattern</code></em> parameter is

  206.         interpreted as a pattern according to the same rules used by

  207.         <span class="application">psql</span>'s <code class="literal">\d</code> commands (see <a class="xref" href="app-psql.html#APP-PSQL-PATTERNS" title="Patterns">Patterns</a>),

  208.         so multiple schemas can also be selected by writing wildcard characters

  209.         in the pattern.  When using wildcards, be careful to quote the pattern

  210.         if needed to prevent the shell from expanding the wildcards;  see

  211.         <a class="xref" href="app-pgdump.html#PG-DUMP-EXAMPLES" title="Examples">Examples</a>.

  212.        </p><div class="note"><h3 class="title">Note</h3><p>

  213.          When <code class="option">-n</code> is specified, <span class="application">pg_dump</span>

  214.          makes no attempt to dump any other database objects that the selected

  215.          schema(s) might depend upon. Therefore, there is no guarantee

  216.          that the results of a specific-schema dump can be successfully

  217.          restored by themselves into a clean database.

  218.         </p></div><div class="note"><h3 class="title">Note</h3><p>

  219.          Non-schema objects such as blobs are not dumped when <code class="option">-n</code> is

  220.          specified.  You can add blobs back to the dump with the

  221.          <code class="option">--blobs</code> switch.

  222.         </p></div></dd><dt><span class="term"><code class="option">-N <em class="replaceable"><code>pattern</code></em></code><br /></span><span class="term"><code class="option">--exclude-schema=<em class="replaceable"><code>pattern</code></em></code></span></dt><dd><p>

  223.         Do not dump any schemas matching <em class="replaceable"><code>pattern</code></em>.  The pattern is

  224.         interpreted according to the same rules as for <code class="option">-n</code>.

  225.         <code class="option">-N</code> can be given more than once to exclude schemas

  226.         matching any of several patterns.

  227.        </p><p>

  228.         When both <code class="option">-n</code> and <code class="option">-N</code> are given, the behavior

  229.         is to dump just the schemas that match at least one <code class="option">-n</code>

  230.         switch but no <code class="option">-N</code> switches.  If <code class="option">-N</code> appears

  231.         without <code class="option">-n</code>, then schemas matching <code class="option">-N</code> are

  232.         excluded from what is otherwise a normal dump.

  233.        </p></dd><dt><span class="term"><code class="option">-O</code><br /></span><span class="term"><code class="option">--no-owner</code></span></dt><dd><p>

  234.         Do not output commands to set

  235.         ownership of objects to match the original database.

  236.         By default, <span class="application">pg_dump</span> issues

  237.         <code class="command">ALTER OWNER</code> or

  238.         <code class="command">SET SESSION AUTHORIZATION</code>

  239.         statements to set ownership of created database objects.

  240.         These statements

  241.         will fail when the script is run unless it is started by a superuser

  242.         (or the same user that owns all of the objects in the script).

  243.         To make a script that can be restored by any user, but will give

  244.         that user ownership of all the objects, specify <code class="option">-O</code>.

  245.        </p><p>

  246.         This option is only meaningful for the plain-text format.  For

  247.         the archive formats, you can specify the option when you

  248.         call <code class="command">pg_restore</code>.

  249.        </p></dd><dt><span class="term"><code class="option">-R</code><br /></span><span class="term"><code class="option">--no-reconnect</code></span></dt><dd><p>

  250.         This option is obsolete but still accepted for backwards

  251.         compatibility.

  252.        </p></dd><dt><span class="term"><code class="option">-s</code><br /></span><span class="term"><code class="option">--schema-only</code></span></dt><dd><p>

  253.         Dump only the object definitions (schema), not data.

  254.        </p><p>

  255.         This option is the inverse of <code class="option">--data-only</code>.

  256.         It is similar to, but for historical reasons not identical to,

  257.         specifying

  258.         <code class="option">--section=pre-data --section=post-data</code>.

  259.        </p><p>

  260.         (Do not confuse this with the <code class="option">--schema</code> option, which

  261.         uses the word <span class="quote">“<span class="quote">schema</span>”</span> in a different meaning.)

  262.        </p><p>

  263.         To exclude table data for only a subset of tables in the database,

  264.         see <code class="option">--exclude-table-data</code>.

  265.        </p></dd><dt><span class="term"><code class="option">-S <em class="replaceable"><code>username</code></em></code><br /></span><span class="term"><code class="option">--superuser=<em class="replaceable"><code>username</code></em></code></span></dt><dd><p>

  266.         Specify the superuser user name to use when disabling triggers.

  267.         This is relevant only if <code class="option">--disable-triggers</code> is used.

  268.         (Usually, it's better to leave this out, and instead start the

  269.         resulting script as superuser.)

  270.        </p></dd><dt><span class="term"><code class="option">-t <em class="replaceable"><code>pattern</code></em></code><br /></span><span class="term"><code class="option">--table=<em class="replaceable"><code>pattern</code></em></code></span></dt><dd><p>

  271.         Dump only tables with names matching

  272.         <em class="replaceable"><code>pattern</code></em>.

  273.         For this purpose, <span class="quote">“<span class="quote">table</span>”</span> includes views, materialized views,

  274.         sequences, and foreign tables.  Multiple tables

  275.         can be selected by writing multiple <code class="option">-t</code> switches.  The

  276.         <em class="replaceable"><code>pattern</code></em> parameter is

  277.         interpreted as a pattern according to the same rules used by

  278.         <span class="application">psql</span>'s <code class="literal">\d</code> commands (see <a class="xref" href="app-psql.html#APP-PSQL-PATTERNS" title="Patterns">Patterns</a>),

  279.         so multiple tables can also be selected by writing wildcard characters

  280.         in the pattern.  When using wildcards, be careful to quote the pattern

  281.         if needed to prevent the shell from expanding the wildcards;  see

  282.         <a class="xref" href="app-pgdump.html#PG-DUMP-EXAMPLES" title="Examples">Examples</a>.

  283.        </p><p>

  284.         The <code class="option">-n</code> and <code class="option">-N</code> switches have no effect when

  285.         <code class="option">-t</code> is used, because tables selected by <code class="option">-t</code> will

  286.         be dumped regardless of those switches, and non-table objects will not

  287.         be dumped.

  288.        </p><div class="note"><h3 class="title">Note</h3><p>

  289.          When <code class="option">-t</code> is specified, <span class="application">pg_dump</span>

  290.          makes no attempt to dump any other database objects that the selected

  291.          table(s) might depend upon. Therefore, there is no guarantee

  292.          that the results of a specific-table dump can be successfully

  293.          restored by themselves into a clean database.

  294.         </p></div><div class="note"><h3 class="title">Note</h3><p>

  295.          The behavior of the <code class="option">-t</code> switch is not entirely upward

  296.          compatible with pre-8.2 <span class="productname">PostgreSQL</span>

  297.          versions.  Formerly, writing <code class="literal">-t tab</code> would dump all

  298.          tables named <code class="literal">tab</code>, but now it just dumps whichever one

  299.          is visible in your default search path.  To get the old behavior

  300.          you can write <code class="literal">-t '*.tab'</code>.  Also, you must write something

  301.          like <code class="literal">-t sch.tab</code> to select a table in a particular schema,

  302.          rather than the old locution of <code class="literal">-n sch -t tab</code>.

  303.         </p></div></dd><dt><span class="term"><code class="option">-T <em class="replaceable"><code>pattern</code></em></code><br /></span><span class="term"><code class="option">--exclude-table=<em class="replaceable"><code>pattern</code></em></code></span></dt><dd><p>

  304.         Do not dump any tables matching <em class="replaceable"><code>pattern</code></em>.  The pattern is

  305.         interpreted according to the same rules as for <code class="option">-t</code>.

  306.         <code class="option">-T</code> can be given more than once to exclude tables

  307.         matching any of several patterns.

  308.        </p><p>

  309.         When both <code class="option">-t</code> and <code class="option">-T</code> are given, the behavior

  310.         is to dump just the tables that match at least one <code class="option">-t</code>

  311.         switch but no <code class="option">-T</code> switches.  If <code class="option">-T</code> appears

  312.         without <code class="option">-t</code>, then tables matching <code class="option">-T</code> are

  313.         excluded from what is otherwise a normal dump.

  314.        </p></dd><dt><span class="term"><code class="option">-v</code><br /></span><span class="term"><code class="option">--verbose</code></span></dt><dd><p>

  315.         Specifies verbose mode.  This will cause

  316.         <span class="application">pg_dump</span> to output detailed object

  317.         comments and start/stop times to the dump file, and progress

  318.         messages to standard error.

  319.        </p></dd><dt><span class="term"><code class="option">-V</code><br /></span><span class="term"><code class="option">--version</code></span></dt><dd><p>

  320.        Print the <span class="application">pg_dump</span> version and exit.

  321.        </p></dd><dt><span class="term"><code class="option">-x</code><br /></span><span class="term"><code class="option">--no-privileges</code><br /></span><span class="term"><code class="option">--no-acl</code></span></dt><dd><p>

  322.         Prevent dumping of access privileges (grant/revoke commands).

  323.        </p></dd><dt><span class="term"><code class="option">-Z <em class="replaceable"><code>0..9</code></em></code><br /></span><span class="term"><code class="option">--compress=<em class="replaceable"><code>0..9</code></em></code></span></dt><dd><p>

  324.         Specify the compression level to use.  Zero means no compression.

  325.         For the custom archive format, this specifies compression of

  326.         individual table-data segments, and the default is to compress

  327.         at a moderate level.

  328.         For plain text output, setting a nonzero compression level causes

  329.         the entire output file to be compressed, as though it had been

  330.         fed through <span class="application">gzip</span>; but the default is not to compress.

  331.         The tar archive format currently does not support compression at all.

  332.        </p></dd><dt><span class="term"><code class="option">--binary-upgrade</code></span></dt><dd><p>

  333.         This option is for use by in-place upgrade utilities.  Its use

  334.         for other purposes is not recommended or supported.  The

  335.         behavior of the option may change in future releases without

  336.         notice.

  337.        </p></dd><dt><span class="term"><code class="option">--column-inserts</code><br /></span><span class="term"><code class="option">--attribute-inserts</code></span></dt><dd><p>

  338.         Dump data as <code class="command">INSERT</code> commands with explicit

  339.         column names (<code class="literal">INSERT INTO

  340.         <em class="replaceable"><code>table</code></em>

  341.         (<em class="replaceable"><code>column</code></em>, ...) VALUES

  342.         ...</code>).  This will make restoration very slow; it is mainly

  343.         useful for making dumps that can be loaded into

  344.         non-<span class="productname">PostgreSQL</span> databases.

  345.         Any error during reloading will cause only rows that are part of the

  346.         problematic <code class="command">INSERT</code> to be lost, rather than the

  347.         entire table contents.

  348.        </p></dd><dt><span class="term"><code class="option">--disable-dollar-quoting</code></span></dt><dd><p>

  349.         This option disables the use of dollar quoting for function bodies,

  350.         and forces them to be quoted using SQL standard string syntax.

  351.        </p></dd><dt><span class="term"><code class="option">--disable-triggers</code></span></dt><dd><p>

  352.         This option is relevant only when creating a data-only dump.

  353.         It instructs <span class="application">pg_dump</span> to include commands

  354.         to temporarily disable triggers on the target tables while

  355.         the data is reloaded.  Use this if you have referential

  356.         integrity checks or other triggers on the tables that you

  357.         do not want to invoke during data reload.

  358.        </p><p>

  359.         Presently, the commands emitted for <code class="option">--disable-triggers</code>

  360.         must be done as superuser.  So, you should also specify

  361.         a superuser name with <code class="option">-S</code>, or preferably be careful to

  362.         start the resulting script as a superuser.

  363.        </p><p>

  364.         This option is only meaningful for the plain-text format.  For

  365.         the archive formats, you can specify the option when you

  366.         call <code class="command">pg_restore</code>.

  367.        </p></dd><dt><span class="term"><code class="option">--enable-row-security</code></span></dt><dd><p>

  368.         This option is relevant only when dumping the contents of a table

  369.         which has row security.  By default, <span class="application">pg_dump</span> will set

  370.         <a class="xref" href="runtime-config-client.html#GUC-ROW-SECURITY">row_security</a> to off, to ensure

  371.         that all data is dumped from the table.  If the user does not have

  372.         sufficient privileges to bypass row security, then an error is thrown.

  373.         This parameter instructs <span class="application">pg_dump</span> to set

  374.         <a class="xref" href="runtime-config-client.html#GUC-ROW-SECURITY">row_security</a> to on instead, allowing the user

  375.         to dump the parts of the contents of the table that they have access to.

  376.        </p><p>

  377.         Note that if you use this option currently, you probably also want

  378.         the dump be in <code class="command">INSERT</code> format, as the

  379.         <code class="command">COPY FROM</code> during restore does not support row security.

  380.        </p></dd><dt><span class="term"><code class="option">--exclude-table-data=<em class="replaceable"><code>pattern</code></em></code></span></dt><dd><p>

  381.         Do not dump data for any tables matching <em class="replaceable"><code>pattern</code></em>. The pattern is

  382.         interpreted according to the same rules as for <code class="option">-t</code>.

  383.         <code class="option">--exclude-table-data</code> can be given more than once to

  384.         exclude tables matching any of several patterns. This option is

  385.         useful when you need the definition of a particular table even

  386.         though you do not need the data in it.

  387.        </p><p>

  388.         To exclude data for all tables in the database, see <code class="option">--schema-only</code>.

  389.        </p></dd><dt><span class="term"><code class="option">--extra-float-digits=<em class="replaceable"><code>ndigits</code></em></code></span></dt><dd><p>

  390.         Use the specified value of <code class="option">extra_float_digits</code> when dumping

  391.         floating-point data, instead of the maximum available precision.

  392.         Routine dumps made for backup purposes should not use this option.

  393.        </p></dd><dt><span class="term"><code class="option">--if-exists</code></span></dt><dd><p>

  394.         Use conditional commands (i.e. add an <code class="literal">IF EXISTS</code>

  395.         clause) when cleaning database objects.  This option is not valid

  396.         unless <code class="option">--clean</code> is also specified.

  397.        </p></dd><dt><span class="term"><code class="option">--inserts</code></span></dt><dd><p>

  398.         Dump data as <code class="command">INSERT</code> commands (rather

  399.         than <code class="command">COPY</code>).  This will make restoration very slow;

  400.         it is mainly useful for making dumps that can be loaded into

  401.         non-<span class="productname">PostgreSQL</span> databases.

  402.         Any error during reloading will cause only rows that are part of the

  403.         problematic <code class="command">INSERT</code> to be lost, rather than the

  404.         entire table contents.  Note that the restore might fail altogether if

  405.         you have rearranged column order.  The

  406.         <code class="option">--column-inserts</code> option is safe against column order

  407.         changes, though even slower.

  408.        </p></dd><dt><span class="term"><code class="option">--load-via-partition-root</code></span></dt><dd><p>

  409.         When dumping data for a table partition, make

  410.         the <code class="command">COPY</code> or <code class="command">INSERT</code> statements

  411.         target the root of the partitioning hierarchy that contains it, rather

  412.         than the partition itself.  This causes the appropriate partition to

  413.         be re-determined for each row when the data is loaded.  This may be

  414.         useful when reloading data on a server where rows do not always fall

  415.         into the same partitions as they did on the original server.  That

  416.         could happen, for example, if the partitioning column is of type text

  417.         and the two systems have different definitions of the collation used

  418.         to sort the partitioning column.

  419.        </p><p>

  420.         It is best not to use parallelism when restoring from an archive made

  421.         with this option, because <span class="application">pg_restore</span> will

  422.         not know exactly which partition(s) a given archive data item will

  423.         load data into.  This could result in inefficiency due to lock

  424.         conflicts between parallel jobs, or perhaps even reload failures due

  425.         to foreign key constraints being set up before all the relevant data

  426.         is loaded.

  427.        </p></dd><dt><span class="term"><code class="option">--lock-wait-timeout=<em class="replaceable"><code>timeout</code></em></code></span></dt><dd><p>

  428.         Do not wait forever to acquire shared table locks at the beginning of

  429.         the dump. Instead fail if unable to lock a table within the specified

  430.         <em class="replaceable"><code>timeout</code></em>. The timeout may be

  431.         specified in any of the formats accepted by <code class="command">SET

  432.         statement_timeout</code>.  (Allowed formats vary depending on the server

  433.         version you are dumping from, but an integer number of milliseconds

  434.         is accepted by all versions.)

  435.        </p></dd><dt><span class="term"><code class="option">--no-comments</code></span></dt><dd><p>

  436.         Do not dump comments.

  437.        </p></dd><dt><span class="term"><code class="option">--no-publications</code></span></dt><dd><p>

  438.         Do not dump publications.

  439.        </p></dd><dt><span class="term"><code class="option">--no-security-labels</code></span></dt><dd><p>

  440.         Do not dump security labels.

  441.        </p></dd><dt><span class="term"><code class="option">--no-subscriptions</code></span></dt><dd><p>

  442.         Do not dump subscriptions.

  443.        </p></dd><dt><span class="term"><code class="option">--no-sync</code></span></dt><dd><p>

  444.         By default, <code class="command">pg_dump</code> will wait for all files

  445.         to be written safely to disk.  This option causes

  446.         <code class="command">pg_dump</code> to return without waiting, which is

  447.         faster, but means that a subsequent operating system crash can leave

  448.         the dump corrupt.  Generally, this option is useful for testing

  449.         but should not be used when dumping data from production installation.

  450.        </p></dd><dt><span class="term"><code class="option">--no-synchronized-snapshots</code></span></dt><dd><p>

  451.         This option allows running <code class="command">pg_dump -j</code> against a pre-9.2

  452.         server, see the documentation of the <code class="option">-j</code> parameter

  453.         for more details.

  454.        </p></dd><dt><span class="term"><code class="option">--no-tablespaces</code></span></dt><dd><p>

  455.         Do not output commands to select tablespaces.

  456.         With this option, all objects will be created in whichever

  457.         tablespace is the default during restore.

  458.        </p><p>

  459.         This option is only meaningful for the plain-text format.  For

  460.         the archive formats, you can specify the option when you

  461.         call <code class="command">pg_restore</code>.

  462.        </p></dd><dt><span class="term"><code class="option">--no-unlogged-table-data</code></span></dt><dd><p>

  463.         Do not dump the contents of unlogged tables.  This option has no

  464.         effect on whether or not the table definitions (schema) are dumped;

  465.         it only suppresses dumping the table data. Data in unlogged tables

  466.         is always excluded when dumping from a standby server.

  467.        </p></dd><dt><span class="term"><code class="option">--on-conflict-do-nothing</code></span></dt><dd><p>

  468.         Add <code class="literal">ON CONFLICT DO NOTHING</code> to

  469.         <code class="command">INSERT</code> commands.

  470.         This option is not valid unless <code class="option">--inserts</code>,

  471.         <code class="option">--column-inserts</code> or

  472.         <code class="option">--rows-per-insert</code> is also specified.

  473.        </p></dd><dt><span class="term"><code class="option">--quote-all-identifiers</code></span></dt><dd><p>

  474.         Force quoting of all identifiers.  This option is recommended when

  475.         dumping a database from a server whose <span class="productname">PostgreSQL</span>

  476.         major version is different from <span class="application">pg_dump</span>'s, or when

  477.         the output is intended to be loaded into a server of a different

  478.         major version.  By default, <span class="application">pg_dump</span> quotes only

  479.         identifiers that are reserved words in its own major version.

  480.         This sometimes results in compatibility issues when dealing with

  481.         servers of other versions that may have slightly different sets

  482.         of reserved words.  Using <code class="option">--quote-all-identifiers</code> prevents

  483.         such issues, at the price of a harder-to-read dump script.

  484.        </p></dd><dt><span class="term"><code class="option">--rows-per-insert=<em class="replaceable"><code>nrows</code></em></code></span></dt><dd><p>

  485.         Dump data as <code class="command">INSERT</code> commands (rather than

  486.         <code class="command">COPY</code>).  Controls the maximum number of rows per

  487.         <code class="command">INSERT</code> command. The value specified must be a

  488.         number greater than zero.  Any error during reloading will cause only

  489.         rows that are part of the problematic <code class="command">INSERT</code> to be

  490.         lost, rather than the entire table contents.

  491.        </p></dd><dt><span class="term"><code class="option">--section=<em class="replaceable"><code>sectionname</code></em></code></span></dt><dd><p>

  492.           Only dump the named section. The section name can be

  493.           <code class="option">pre-data</code>, <code class="option">data</code>, or <code class="option">post-data</code>.

  494.           This option can be specified more than once to select multiple

  495.           sections. The default is to dump all sections.

  496.          </p><p>

  497.           The data section contains actual table data, large-object

  498.           contents, and sequence values.

  499.           Post-data items include definitions of indexes, triggers, rules,

  500.           and constraints other than validated check constraints.

  501.           Pre-data items include all other data definition items.

  502.          </p></dd><dt><span class="term"><code class="option">--serializable-deferrable</code></span></dt><dd><p>

  503.         Use a <code class="literal">serializable</code> transaction for the dump, to

  504.         ensure that the snapshot used is consistent with later database

  505.         states; but do this by waiting for a point in the transaction stream

  506.         at which no anomalies can be present, so that there isn't a risk of

  507.         the dump failing or causing other transactions to roll back with a

  508.         <code class="literal">serialization_failure</code>.  See <a class="xref" href="mvcc.html" title="Chapter 13. Concurrency Control">Chapter 13</a>

  509.         for more information about transaction isolation and concurrency

  510.         control.

  511.        </p><p>

  512.         This option is not beneficial for a dump which is intended only for

  513.         disaster recovery.  It could be useful for a dump used to load a

  514.         copy of the database for reporting or other read-only load sharing

  515.         while the original database continues to be updated.  Without it the

  516.         dump may reflect a state which is not consistent with any serial

  517.         execution of the transactions eventually committed.  For example, if

  518.         batch processing techniques are used, a batch may show as closed in

  519.         the dump without all of the items which are in the batch appearing.

  520.        </p><p>

  521.         This option will make no difference if there are no read-write

  522.         transactions active when pg_dump is started.  If read-write

  523.         transactions are active, the start of the dump may be delayed for an

  524.         indeterminate length of time.  Once running, performance with or

  525.         without the switch is the same.

  526.        </p></dd><dt><span class="term"><code class="option">--snapshot=<em class="replaceable"><code>snapshotname</code></em></code></span></dt><dd><p>

  527.           Use the specified synchronized snapshot when making a dump of the

  528.           database (see

  529.           <a class="xref" href="functions-admin.html#FUNCTIONS-SNAPSHOT-SYNCHRONIZATION-TABLE" title="Table 9.87. Snapshot Synchronization Functions">Table 9.87</a> for more

  530.           details).

  531.          </p><p>

  532.           This option is useful when needing to synchronize the dump with

  533.           a logical replication slot (see <a class="xref" href="logicaldecoding.html" title="Chapter 48. Logical Decoding">Chapter 48</a>)

  534.           or with a concurrent session.

  535.          </p><p>

  536.           In the case of a parallel dump, the snapshot name defined by this

  537.           option is used rather than taking a new snapshot.

  538.          </p></dd><dt><span class="term"><code class="option">--strict-names</code></span></dt><dd><p>

  539.         Require that each schema

  540.         (<code class="option">-n</code>/<code class="option">--schema</code>) and table

  541.         (<code class="option">-t</code>/<code class="option">--table</code>) qualifier match at

  542.         least one schema/table in the database to be dumped.  Note that if

  543.         none of the schema/table qualifiers find

  544.         matches, <span class="application">pg_dump</span> will generate an error

  545.         even without <code class="option">--strict-names</code>.

  546.        </p><p>

  547.         This option has no effect

  548.         on <code class="option">-N</code>/<code class="option">--exclude-schema</code>,

  549.         <code class="option">-T</code>/<code class="option">--exclude-table</code>,

  550.         or <code class="option">--exclude-table-data</code>.  An exclude pattern failing

  551.         to match any objects is not considered an error.

  552.        </p></dd><dt><span class="term"><code class="option">--use-set-session-authorization</code></span></dt><dd><p>

  553.         Output SQL-standard <code class="command">SET SESSION AUTHORIZATION</code> commands

  554.         instead of <code class="command">ALTER OWNER</code> commands to determine object

  555.         ownership.  This makes the dump more standards-compatible, but

  556.         depending on the history of the objects in the dump, might not restore

  557.         properly.  Also, a dump using <code class="command">SET SESSION AUTHORIZATION</code>

  558.         will certainly require superuser privileges to restore correctly,

  559.         whereas <code class="command">ALTER OWNER</code> requires lesser privileges.

  560.        </p></dd><dt><span class="term"><code class="option">-?</code><br /></span><span class="term"><code class="option">--help</code></span></dt><dd><p>

  561.        Show help about <span class="application">pg_dump</span> command line

  562.        arguments, and exit.

  563.        </p></dd></dl></div><p>

  564.    </p><p>

  565.     The following command-line options control the database connection parameters.

  566. 

  567.     </p><div class="variablelist"><dl class="variablelist"><dt><span class="term"><code class="option">-d <em class="replaceable"><code>dbname</code></em></code><br /></span><span class="term"><code class="option">--dbname=<em class="replaceable"><code>dbname</code></em></code></span></dt><dd><p>

  568.        Specifies the name of the database to connect to. This is

  569.        equivalent to specifying <em class="replaceable"><code>dbname</code></em> as the first non-option

  570.        argument on the command line.

  571.       </p><p>

  572.        If this parameter contains an <code class="symbol">=</code> sign or starts

  573.        with a valid <acronym class="acronym">URI</acronym> prefix

  574.        (<code class="literal">postgresql://</code>

  575.        or <code class="literal">postgres://</code>), it is treated as a

  576.        <em class="parameter"><code>conninfo</code></em> string. See <a class="xref" href="libpq-connect.html" title="33.1. Database Connection Control Functions">Section 33.1</a> for more information.

  577.       </p></dd><dt><span class="term"><code class="option">-h <em class="replaceable"><code>host</code></em></code><br /></span><span class="term"><code class="option">--host=<em class="replaceable"><code>host</code></em></code></span></dt><dd><p>

  578.         Specifies the host name of the machine on which the server is

  579.         running.  If the value begins with a slash, it is used as the

  580.         directory for the Unix domain socket. The default is taken

  581.         from the <code class="envar">PGHOST</code> environment variable, if set,

  582.         else a Unix domain socket connection is attempted.

  583.        </p></dd><dt><span class="term"><code class="option">-p <em class="replaceable"><code>port</code></em></code><br /></span><span class="term"><code class="option">--port=<em class="replaceable"><code>port</code></em></code></span></dt><dd><p>

  584.         Specifies the TCP port or local Unix domain socket file

  585.         extension on which the server is listening for connections.

  586.         Defaults to the <code class="envar">PGPORT</code> environment variable, if

  587.         set, or a compiled-in default.

  588.        </p></dd><dt><span class="term"><code class="option">-U <em class="replaceable"><code>username</code></em></code><br /></span><span class="term"><code class="option">--username=<em class="replaceable"><code>username</code></em></code></span></dt><dd><p>

  589.         User name to connect as.

  590.        </p></dd><dt><span class="term"><code class="option">-w</code><br /></span><span class="term"><code class="option">--no-password</code></span></dt><dd><p>

  591.         Never issue a password prompt.  If the server requires

  592.         password authentication and a password is not available by

  593.         other means such as a <code class="filename">.pgpass</code> file, the

  594.         connection attempt will fail.  This option can be useful in

  595.         batch jobs and scripts where no user is present to enter a

  596.         password.

  597.        </p></dd><dt><span class="term"><code class="option">-W</code><br /></span><span class="term"><code class="option">--password</code></span></dt><dd><p>

  598.         Force <span class="application">pg_dump</span> to prompt for a

  599.         password before connecting to a database.

  600.        </p><p>

  601.         This option is never essential, since

  602.         <span class="application">pg_dump</span> will automatically prompt

  603.         for a password if the server demands password authentication.

  604.         However, <span class="application">pg_dump</span> will waste a

  605.         connection attempt finding out that the server wants a password.

  606.         In some cases it is worth typing <code class="option">-W</code> to avoid the extra

  607.         connection attempt.

  608.        </p></dd><dt><span class="term"><code class="option">--role=<em class="replaceable"><code>rolename</code></em></code></span></dt><dd><p>

  609.         Specifies a role name to be used to create the dump.

  610.         This option causes <span class="application">pg_dump</span> to issue a

  611.         <code class="command">SET ROLE</code> <em class="replaceable"><code>rolename</code></em>

  612.         command after connecting to the database. It is useful when the

  613.         authenticated user (specified by <code class="option">-U</code>) lacks privileges

  614.         needed by <span class="application">pg_dump</span>, but can switch to a role with

  615.         the required rights.  Some installations have a policy against

  616.         logging in directly as a superuser, and use of this option allows

  617.         dumps to be made without violating the policy.

  618.        </p></dd></dl></div><p>

  619.    </p></div><div class="refsect1" id="id-1.9.4.12.7"><h2>Environment</h2><div class="variablelist"><dl class="variablelist"><dt><span class="term"><code class="envar">PGDATABASE</code><br /></span><span class="term"><code class="envar">PGHOST</code><br /></span><span class="term"><code class="envar">PGOPTIONS</code><br /></span><span class="term"><code class="envar">PGPORT</code><br /></span><span class="term"><code class="envar">PGUSER</code></span></dt><dd><p>

  620.       Default connection parameters.

  621.      </p></dd><dt><span class="term"><code class="envar">PG_COLOR</code></span></dt><dd><p>

  622.       Specifies whether to use color in diagnostic messages. Possible values

  623.       are <code class="literal">always</code>, <code class="literal">auto</code> and

  624.       <code class="literal">never</code>.

  625.      </p></dd></dl></div><p>

  626.    This utility, like most other <span class="productname">PostgreSQL</span> utilities,

  627.    also uses the environment variables supported by <span class="application">libpq</span>

  628.    (see <a class="xref" href="libpq-envars.html" title="33.14. Environment Variables">Section 33.14</a>).

  629.   </p></div><div class="refsect1" id="APP-PGDUMP-DIAGNOSTICS"><h2>Diagnostics</h2><p>

  630.    <span class="application">pg_dump</span> internally executes

  631.    <code class="command">SELECT</code> statements. If you have problems running

  632.    <span class="application">pg_dump</span>, make sure you are able to

  633.    select information from the database using, for example, <a class="xref" href="app-psql.html" title="psql"><span class="refentrytitle"><span class="application">psql</span></span></a>.  Also, any default connection settings and environment

  634.    variables used by the <span class="application">libpq</span> front-end

  635.    library will apply.

  636.   </p><p>

  637.    The database activity of <span class="application">pg_dump</span> is

  638.    normally collected by the statistics collector.  If this is

  639.    undesirable, you can set parameter <code class="varname">track_counts</code>

  640.    to false via <code class="envar">PGOPTIONS</code> or the <code class="literal">ALTER

  641.    USER</code> command.

  642.   </p></div><div class="refsect1" id="PG-DUMP-NOTES"><h2>Notes</h2><p>

  643.    If your database cluster has any local additions to the <code class="literal">template1</code> database,

  644.    be careful to restore the output of <span class="application">pg_dump</span> into a

  645.    truly empty database; otherwise you are likely to get errors due to

  646.    duplicate definitions of the added objects.  To make an empty database

  647.    without any local additions, copy from <code class="literal">template0</code> not <code class="literal">template1</code>,

  648.    for example:

  649. </p><pre class="programlisting">

  650. CREATE DATABASE foo WITH TEMPLATE template0;

  651. </pre><p>

  652.   </p><p>

  653.    When a data-only dump is chosen and the option <code class="option">--disable-triggers</code>

  654.    is used, <span class="application">pg_dump</span> emits commands

  655.    to disable triggers on user tables before inserting the data,

  656.    and then commands to re-enable them after the data has been

  657.    inserted.  If the restore is stopped in the middle, the system

  658.    catalogs might be left in the wrong state.

  659.   </p><p>

  660.    The dump file produced by <span class="application">pg_dump</span>

  661.    does not contain the statistics used by the optimizer to make

  662.    query planning decisions.  Therefore, it is wise to run

  663.    <code class="command">ANALYZE</code> after restoring from a dump file

  664.    to ensure optimal performance; see <a class="xref" href="routine-vacuuming.html#VACUUM-FOR-STATISTICS" title="24.1.3. Updating Planner Statistics">Section 24.1.3</a>

  665.    and <a class="xref" href="routine-vacuuming.html#AUTOVACUUM" title="24.1.6. The Autovacuum Daemon">Section 24.1.6</a> for more information.

  666.   </p><p>

  667.    Because <span class="application">pg_dump</span> is used to transfer data

  668.    to newer versions of <span class="productname">PostgreSQL</span>, the output of

  669.    <span class="application">pg_dump</span> can be expected to load into

  670.    <span class="productname">PostgreSQL</span> server versions newer than

  671.    <span class="application">pg_dump</span>'s version.  <span class="application">pg_dump</span> can also

  672.    dump from <span class="productname">PostgreSQL</span> servers older than its own version.

  673.    (Currently, servers back to version 8.0 are supported.)

  674.    However, <span class="application">pg_dump</span> cannot dump from

  675.    <span class="productname">PostgreSQL</span> servers newer than its own major version;

  676.    it will refuse to even try, rather than risk making an invalid dump.

  677.    Also, it is not guaranteed that <span class="application">pg_dump</span>'s output can

  678.    be loaded into a server of an older major version — not even if the

  679.    dump was taken from a server of that version.  Loading a dump file

  680.    into an older server may require manual editing of the dump file

  681.    to remove syntax not understood by the older server.

  682.    Use of the <code class="option">--quote-all-identifiers</code> option is recommended

  683.    in cross-version cases, as it can prevent problems arising from varying

  684.    reserved-word lists in different <span class="productname">PostgreSQL</span> versions.

  685.   </p><p>

  686.    When dumping logical replication subscriptions,

  687.    <span class="application">pg_dump</span> will generate <code class="command">CREATE

  688.    SUBSCRIPTION</code> commands that use the <code class="literal">connect = false</code>

  689.    option, so that restoring the subscription does not make remote connections

  690.    for creating a replication slot or for initial table copy.  That way, the

  691.    dump can be restored without requiring network access to the remote

  692.    servers.  It is then up to the user to reactivate the subscriptions in a

  693.    suitable way.  If the involved hosts have changed, the connection

  694.    information might have to be changed.  It might also be appropriate to

  695.    truncate the target tables before initiating a new full table copy.

  696.   </p></div><div class="refsect1" id="PG-DUMP-EXAMPLES"><h2>Examples</h2><p>

  697.    To dump a database called <code class="literal">mydb</code> into a SQL-script file:

  698. </p><pre class="screen">

  699. <code class="prompt">$</code> <strong class="userinput"><code>pg_dump mydb &gt; db.sql</code></strong>

  700. </pre><p>

  701.   </p><p>

  702.    To reload such a script into a (freshly created) database named

  703.    <code class="literal">newdb</code>:

  704. 

  705. </p><pre class="screen">

  706. <code class="prompt">$</code> <strong class="userinput"><code>psql -d newdb -f db.sql</code></strong>

  707. </pre><p>

  708.   </p><p>

  709.    To dump a database into a custom-format archive file:

  710. 

  711. </p><pre class="screen">

  712. <code class="prompt">$</code> <strong class="userinput"><code>pg_dump -Fc mydb &gt; db.dump</code></strong>

  713. </pre><p>

  714.   </p><p>

  715.    To dump a database into a directory-format archive:

  716. 

  717. </p><pre class="screen">

  718. <code class="prompt">$</code> <strong class="userinput"><code>pg_dump -Fd mydb -f dumpdir</code></strong>

  719. </pre><p>

  720.   </p><p>

  721.    To dump a database into a directory-format archive in parallel with

  722.    5 worker jobs:

  723. 

  724. </p><pre class="screen">

  725. <code class="prompt">$</code> <strong class="userinput"><code>pg_dump -Fd mydb -j 5 -f dumpdir</code></strong>

  726. </pre><p>

  727.   </p><p>

  728.    To reload an archive file into a (freshly created) database named

  729.    <code class="literal">newdb</code>:

  730. 

  731. </p><pre class="screen">

  732. <code class="prompt">$</code> <strong class="userinput"><code>pg_restore -d newdb db.dump</code></strong>

  733. </pre><p>

  734.   </p><p>

  735.    To reload an archive file into the same database it was dumped from,

  736.    discarding the current contents of that database:

  737. 

  738. </p><pre class="screen">

  739. <code class="prompt">$</code> <strong class="userinput"><code>pg_restore -d postgres --clean --create db.dump</code></strong>

  740. </pre><p>

  741.   </p><p>

  742.    To dump a single table named <code class="literal">mytab</code>:

  743. 

  744. </p><pre class="screen">

  745. <code class="prompt">$</code> <strong class="userinput"><code>pg_dump -t mytab mydb &gt; db.sql</code></strong>

  746. </pre><p>

  747.   </p><p>

  748.    To dump all tables whose names start with <code class="literal">emp</code> in the

  749.    <code class="literal">detroit</code> schema, except for the table named

  750.    <code class="literal">employee_log</code>:

  751. 

  752. </p><pre class="screen">

  753. <code class="prompt">$</code> <strong class="userinput"><code>pg_dump -t 'detroit.emp*' -T detroit.employee_log mydb &gt; db.sql</code></strong>

  754. </pre><p>

  755.   </p><p>

  756.    To dump all schemas whose names start with <code class="literal">east</code> or

  757.    <code class="literal">west</code> and end in <code class="literal">gsm</code>, excluding any schemas whose

  758.    names contain the word <code class="literal">test</code>:

  759. 

  760. </p><pre class="screen">

  761. <code class="prompt">$</code> <strong class="userinput"><code>pg_dump -n 'east*gsm' -n 'west*gsm' -N '*test*' mydb &gt; db.sql</code></strong>

  762. </pre><p>

  763.   </p><p>

  764.    The same, using regular expression notation to consolidate the switches:

  765. 

  766. </p><pre class="screen">

  767. <code class="prompt">$</code> <strong class="userinput"><code>pg_dump -n '(east|west)*gsm' -N '*test*' mydb &gt; db.sql</code></strong>

  768. </pre><p>

  769.   </p><p>

  770.    To dump all database objects except for tables whose names begin with

  771.    <code class="literal">ts_</code>:

  772. 

  773. </p><pre class="screen">

  774. <code class="prompt">$</code> <strong class="userinput"><code>pg_dump -T 'ts_*' mydb &gt; db.sql</code></strong>

  775. </pre><p>

  776.   </p><p>

  777.    To specify an upper-case or mixed-case name in <code class="option">-t</code> and related

  778.    switches, you need to double-quote the name; else it will be folded to

  779.    lower case (see <a class="xref" href="app-psql.html#APP-PSQL-PATTERNS" title="Patterns">Patterns</a>).  But

  780.    double quotes are special to the shell, so in turn they must be quoted.

  781.    Thus, to dump a single table with a mixed-case name, you need something

  782.    like

  783. 

  784. </p><pre class="screen">

  785. <code class="prompt">$</code> <strong class="userinput"><code>pg_dump -t "\"MixedCaseName\"" mydb &gt; mytab.sql</code></strong>

  786. </pre></div><div class="refsect1" id="id-1.9.4.12.11"><h2>See Also</h2><span class="simplelist"><a class="xref" href="app-pg-dumpall.html" title="pg_dumpall"><span class="refentrytitle"><span class="application">pg_dumpall</span></span></a>, <a class="xref" href="app-pgrestore.html" title="pg_restore"><span class="refentrytitle">pg_restore</span></a>, <a class="xref" href="app-psql.html" title="psql"><span class="refentrytitle"><span class="application">psql</span></span></a></span></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="app-pgconfig.html">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="reference-client.html">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="app-pg-dumpall.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">pg_config </td><td width="20%" align="center"><a accesskey="h" href="index.html">Home</a></td><td width="40%" align="right" valign="top"> <span class="application">pg_dumpall</span></td></tr></table></div></body></html>
上海开阖软件有限公司 沪ICP备12045867号-1