postgresql/doc/src/sgml/catalogs.sgml

<!--
 Documentation of the system catalogs, directed toward PostgreSQL developers
 $Header: /cvsroot/pgsql/doc/src/sgml/catalogs.sgml,v 2.14 2001/03/24 23:03:26 petere Exp $
 -->

<chapter id="catalogs">
 <title>System Catalogs</title>

 <section id="catalogs-overview">
  <title>Overview</title>

  <para>
   The system catalogs are the place where a relational database
   management system stores schema metadata, such as information about
   tables and columns, and internal bookkeeping information.
   <productname>PostgreSQL</productname>'s system catalogs are regular
   tables.  You can drop and recreate the tables, add columns, insert
   and update values, and severely mess up your system that way.
   Normally one never has to change the system catalogs by hand, there
   are always SQL commands to do that.  (For example, <command>CREATE
   DATABASE</command> inserts a row into the
   <structname>pg_database</structname> catalog -- and actually
   creates the database on disk.)  There are some exceptions for
   esoteric operations, such as adding index access methods.
  </para>

  <table>
   <title>System Catalogs</title>

   <tgroup cols="2">
    <thead>
     <row>
      <entry>Catalog Name</entry>
      <entry>Purpose</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>pg_aggregate</entry>
      <entry>aggregate functions</entry>
     </row>

     <row>
      <entry>pg_am</entry>
      <entry>index access methods</entry>
     </row>

     <row>
      <entry>pg_amop</entry>
      <entry>access method operators</entry>
     </row>

     <row>
      <entry>pg_amproc</entry>
      <entry>access method support procedures</entry>
     </row>

     <row>
      <entry>pg_attrdef</entry>
      <entry>column default values</entry>
     </row>

     <row>
      <entry>pg_attribute</entry>
      <entry>table columns (attributes, fields)</entry>
     </row>

     <row>
      <entry>pg_class</entry>
      <entry>tables, indexes, sequences (<quote>relations</quote>)</entry>
     </row>

     <row>
      <entry>pg_database</entry>
      <entry>databases</entry>
     </row>

     <row>
      <entry>pg_description</entry>
      <entry>descriptions or comments on database objects</entry>
     </row>

     <row>
      <entry>pg_group</entry>
      <entry>user groups</entry>
     </row>

     <row>
      <entry>pg_index</entry>
      <entry>additional index information</entry>
     </row>

     <row>
      <entry>pg_inheritproc</entry>
      <entry>(not used)</entry>
     </row>

     <row>
      <entry>pg_inherits</entry>
      <entry>table inheritance hierarchy</entry>
     </row>

     <row>
      <entry>pg_ipl</entry>
      <entry>(not used)</entry>
     </row>

     <row>
      <entry>pg_language</entry>
      <entry>languages for writing functions</entry>
     </row>

     <row>
      <entry>pg_largeobject</entry>
      <entry>large objects</entry>
     </row>

     <row>
      <entry>pg_listener</entry>
      <entry>asynchronous notification</entry>
     </row>

     <row>
      <entry>pg_opclass</entry>
      <entry>index access method operator classes</entry>
     </row>

     <row>
      <entry>pg_operator</entry>
      <entry>operators</entry>
     </row>

     <row>
      <entry>pg_proc</entry>
      <entry>functions and procedures</entry>
     </row>

     <row>
      <entry>pg_relcheck</entry>
      <entry>check constraints</entry>
     </row>

     <row>
      <entry>pg_rewrite</entry>
      <entry>query rewriter rules</entry>
     </row>

     <row>
      <entry>pg_shadow</entry>
      <entry>database users</entry>
     </row>

     <row>
      <entry>pg_statistic</entry>
      <entry>optimizer statistics</entry>
     </row>

     <row>
      <entry>pg_trigger</entry>
      <entry>triggers</entry>
     </row>

     <row>
      <entry>pg_type</entry>
      <entry>data types</entry>
     </row>
    </tbody>
   </tgroup>
  </table>

  <para>
   More detailed documentation of most catalogs follow below.  The
   catalogs that relate to index access methods are explained in the
   <citetitle>Programmer's Guide</citetitle>.  Some catalogs don't
   have any documentation, yet.
  </para>

 </section>


 <section id="catalog-pg-aggregate">
  <title>pg_aggregate</title>

  <para>
   <structname>pg_aggregate</structname> stores information about
   aggregate functions.  An aggregate function is a function that
   operates on a set of values (typically one column from each the row
   that matches a query condition) and returns a single value computed
   from all these values.  Typical aggregate functions are
   <function>sum</function>, <function>count</function>, and
   <function>max</function>.
  </para>

  <table>
   <title>pg_aggregate Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>
    <tbody>
     <row>
      <entry>aggname</entry>
      <entry><type>name</type></entry>
      <entry></entry>
      <entry>Name of the aggregate function</entry>
     </row>
     <row>
      <entry>aggowner</entry>
      <entry><type>int4</type></entry>
      <entry>pg_shadow.usesysid</entry>
      <entry>Owner (creator) of the aggregate function</entry>
     </row>
     <row>
      <entry>aggtransfn</entry>
      <entry><type>regproc</type> (function)</entry>
      <entry></entry>
      <entry>Transition function</entry>
     </row>
     <row>
      <entry>aggfinalfn</entry>
      <entry><type>regproc</type> (function)</entry>
      <entry></entry>
      <entry>Final function</entry>
     </row>
     <row>
      <entry>aggbasetype</entry>
      <entry><type>oid</type></entry>
      <entry>pg_type.oid</entry>
      <entry>The type on which this function operates when invoked from SQL</entry>
     </row>
     <row>
      <entry>aggtranstype</entry>
      <entry><type>oid</type></entry>
      <entry>pg_type.oid</entry>
      <entry>The type of the aggregate function's internal transition (state) data</entry>
     </row>
     <row>
      <entry>aggfinaltype</entry>
      <entry><type>oid</type></entry>
      <entry>pg_type.oid</entry>
      <entry>The type of the result</entry>
     </row>
     <row>
      <entry>agginitval</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>
       The initial value of the transition state.  This is a text
       field which will be cast to the type of aggtranstype.
      </entry>
     </row>
    </tbody>
   </tgroup>
  </table>

  <para>
   New aggregate functions are registered with the <command>CREATE
   AGGREGATE</command> command.  See the <citetitle>Programmer's
   Guide</citetitle> for more information about writing aggregate
   functions and the meaning of the transition functions, etc.
  </para>

  <para>
   An aggregate function is identified through name
   <emphasis>and</emphasis> argument type.  Hence aggname and aggname
   are the composite primary key.
  </para>

 </section>


 <section id="catalog-pg-attrdef">
  <title>pg_attrdef</title>

  <para>
   This catalog stores column default values.  The main information
   about columns is stored in <structname>pg_attribute</structname>
   (see below).  Only columns that explicitly specify a default value
   (when the table is created or the column is added) will have an
   entry here.
  </para>

  <table>
   <title>pg_attrdef Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>adrelid</entry>
      <entry><type>oid</type></entry>
      <entry>pg_class.oid</entry>
      <entry>The table this column belongs to</entry>
     </row>

     <row>
      <entry>adnum</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>
       The number of the column; see
       <structname>pg_attribute</structname>.<structfield>pg_attnum</structfield>
      </entry>
     </row>

     <row>
      <entry>adbin</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>An internal representation of the column default value</entry>
     </row>

     <row>
      <entry>adsrc</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>A human-readable representation of the default value</entry>
     </row>
    </tbody>
   </tgroup>
  </table>

 </section>


 <section id="catalog-pg-attribute">
  <title>pg_attribute</title>

  <para>
   <structname>pg_attribute</structname> stores information about
   table columns.  There will be exactly one
   <structname>pg_attribute</structname> row for every column in every
   table in the database.  (There will also be attribute entries for
   indexes and other objects.  See <structname>pg_class</structname>.)
  </para>

  <para>
   The term attribute is equivalent to column and is used for
   historical reasons.
  </para>

  <table>
   <title>pg_attribute Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>attrelid</entry>
      <entry><type>oid</type></entry>
      <entry>pg_class.oid</entry>
      <entry>The table this column belongs to</entry>
     </row>

     <row>
      <entry>attname</entry>
      <entry><type>name</type></entry>
      <entry></entry>
      <entry>Column name</entry>
     </row>

     <row>
      <entry>atttypid</entry>
      <entry><type>oid</type></entry>
      <entry>pg_type.oid</entry>
      <entry>The data type of this column</entry>
     </row>

     <row>
      <entry>attdispersion</entry>
      <entry><type>float4</type></entry>
      <entry></entry>
      <entry>
       <structfield>attdispersion</structfield> is the dispersion
       statistic of the column (0.0 to 1.0), or zero if the statistic
       has not been calculated, or -1.0 if <command>VACUUM</command>
       found that the column contains no duplicate entries (in which
       case the dispersion should be taken as
       1.0/<symbol>numberOfRows</symbol> for the current table size).
       The -1.0 hack is useful because the number of rows may be
       updated more often than
       <structfield>attdispersion</structfield> is. We assume that the
       column will retain its no-duplicate-entry property.
      </entry>
     </row>

     <row>
      <entry>attlen</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>
       This is a copy of the
       <structname>pg_type</structname>.<structfield>typlen</structfield>
       for this column's type.
      </entry>
     </row>

     <row>
      <entry>attnum</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>
       The number of the column.  Ordinary columns are numbered from 1
       up.  System columns, such as <structfield>oid</structfield>,
       have (arbitrary) negative numbers.
      </entry>
     </row>

     <row>
      <entry>attnelems</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>Number of dimensions, if the column is an array</entry>
     </row>

     <row>
      <entry>attcacheoff</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>
       Always -1 in storage, but when loaded into a tuple descriptor
       in memory this may be updated cache the offset of the attribute
       within the tuple.
      </entry>
     </row>

     <row>
      <entry>atttypmod</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>
       <structfield>atttypmod</structfield> records type-specific data
       supplied at table creation time (for example, the maximum
       length of a <type>varchar</type> column).  It is passed to
       type-specific input and output functions as the third
       argument. The value will generally be -1 for types that do not
       need typmod.
      </entry>
     </row>

     <row>
      <entry>attbyval</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       A copy of
       <structname>pg_type</structname>.<structfield>typbyval</structfield>
       of this column's type
      </entry>
     </row>

     <row>
      <entry>attstorage</entry>
      <entry><type>char</type></entry>
      <entry></entry>
      <entry>
       A copy of
       <structname>pg_type</structname>.<structfield>typstorage</structfield>
       of this column's type
      </entry>
     </row>

     <row>
      <entry>attisset</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       If true, this attribute is a set.  In that case, what is really
       stored in the attribute is the OID of a tuple in the
       <structname>pg_proc</structname> catalog.  The
       <structname>pg_proc</structname> tuple contains the query
       string that defines this set - i.e., the query to run to get
       the set.  So the <structfield>atttypid</structfield> (see
       above) refers to the type returned by this query, but the
       actual length of this attribute is the length (size) of an
       <type>oid</type>.  --- At least this is the theory.  All this
       is probably quite broken these days.
      </entry>
     </row>

     <row>
      <entry>attalign</entry>
      <entry><type>char</type></entry>
      <entry></entry>
      <entry>
       A copy of
       <structname>pg_type</structname>.<structfield>typalign</structfield>
       of this column's type
      </entry>
     </row>

     <row>
      <entry>attnotnull</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       This represents a NOT NULL constraint.  It is possible to
       change this field to enable or disable the constraint.
      </entry>
     </row>

     <row>
      <entry>atthasdef</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       This column has a default value, in which case there will be a
       corresponding entry in the <structname>pg_attrdef</structname>
       catalog that actually defines the value.
      </entry>
     </row>

    </tbody>
   </tgroup>
  </table>
 </section>


 <section id="catalog-pg-class">
  <title>pg_class</title>

  <para>
   <structname>pg_class</structname> catalogues tables and mostly
   everything else that has columns or is otherwise similar to a
   table.  This includes indexes (but see
   <structname>pg_index</structname>), sequences, views, and some
   kinds of special relation kinds.  Below, when we mean all of these
   kinds of objects we speak of <quote>relations</quote>.  Not all
   fields are meaningful for all relation types.
  </para>

  <table>
   <title>pg_class Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>relname</entry>
      <entry><type>name</type></entry>
      <entry></entry>
      <entry>Name of the table, index, view, etc.</entry>
     </row>

     <row>
      <entry>reltype</entry>
      <entry><type>oid</type></entry>
      <entry>pg_type.oid</entry>
      <entry>
       The data type that corresponds to this table (not functional,
       only set for system tables)
      </entry>
     </row>

     <row>
      <entry>relowner</entry>
      <entry><type>int4</type></entry>
      <entry>pg_shadow.usesysid</entry>
      <entry>Owner of the relation</entry>
     </row>

     <row>
      <entry>relam</entry>
      <entry><type>oid</type></entry>
      <entry>pg_am.oid</entry>
      <entry>If this is an index, the access method used (btree, hash, etc.)</entry>
     </row>

     <row>
      <entry>relfilenode</entry>
      <entry><type>oid</type></entry>
      <entry></entry>
      <entry>Name of the on-disk file of this relation</entry>
     </row>

     <row>
      <entry>relpages</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>
       Size of the on-disk representation of this table in pages (size
       <symbol>BLCKSZ</symbol>).  This is only an approximate value
       which is calculated during vacuum.
      </entry>
     </row>

     <row>
      <entry>reltuples</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>
       Number of tuples in the table.  This is only an estimate used
       by the planner, updated by <command>VACUUM</command>.
      </entry>
     </row>

     <row>
      <entry>reltoastrelid</entry>
      <entry><type>oid</type></entry>
      <entry>pg_class.oid</entry>
      <entry>
       Oid of the TOAST table associated with this table, 0 if none.
       The TOAST table stores large attributes <quote>out of
       line</quote> in a secondary table.
      </entry>
     </row>

     <row>
      <entry>reltoastidxid</entry>
      <entry><type>oid</type></entry>
      <entry>pg_class.oid</entry>
      <entry>Oid of the index on the TOAST table for this table, 0 if none</entry>
     </row>

     <row>
      <entry>relhasindex</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>True if this is a table and it has at least one index</entry>
     </row>

     <row>
      <entry>relisshared</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>XXX (This is not what it seems to be.)</entry>
     </row>

     <row>
      <entry>relkind</entry>
      <entry><type>char</type></entry>
      <entry></entry>
      <entry>
       'r' = ordinary table, 'i' = index, 'S' = sequence, 'v' = view,
       's' = special, 't' = secondary TOAST table
      </entry>
     </row>

     <row>
      <entry>relnatts</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>
       Number of columns in the relation, besides system columns.
       There must be this many corresponding entries in
       <structname>pg_attribute</structname>.  See also
       <structname>pg_attribute</structname>.<structfield>attnum</structfield>.
      </entry>
     </row>

     <row>
      <entry>relchecks</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>
       Number of check constraints on the table; see
       <structname>pg_relcheck</structname> catalog
      </entry>
     </row>

     <row>
      <entry>reltriggers</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>
       Number of triggers on the table; see
       <structname>pg_trigger</structname> catalog
      </entry>
     </row>

     <row>
      <entry>relukeys</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>unused  (<emphasis>Not</emphasis> the number of unique keys or something.)</entry>
     </row>

     <row>
      <entry>relfkeys</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>Number foreign keys on the table</entry>
     </row>

     <row>
      <entry>relhaspkey</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       unused  (No, this does not say whether the table has a primary
       key.  It's really unused.)
      </entry>
     </row>

     <row>
      <entry>relhasrules</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>Table has rules</entry>
     </row>

     <row>
      <entry>relhassubclass</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>At least one table inherits this one</entry>
     </row>

     <row>
      <entry>relacl</entry>
      <entry><type>aclitem[]</type></entry>
      <entry></entry>
      <entry>
       Access permissions.  See the descriptions of
       <command>GRANT</command> and <command>REVOKE</command> for
       details.
      </entry>
     </row>
    </tbody>
   </tgroup>
  </table>
 </section>


 <section id="catalog-pg-database">
  <title>pg_database</title>

  <para>
   The <structname>pg_database</structname> catalog stores information
   about the available databases.  The
   <structname>pg_database</structname> table is shared between all
   databases of a cluster.  Databases are created with the
   <command>CREATE DATABASE</command>.  Consult the
   <citetitle>Administrator's Guide</citetitle> for details about the
   meaning of some of the parameters.
  </para>

  <table>
   <title>pg_database Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>datname</entry>
      <entry><type>name</type></entry>
      <entry></entry>
      <entry>Database name</entry>
     </row>

     <row>
      <entry>datdba</entry>
      <entry><type>int4</type></entry>
      <entry>pg_shadow.usesysid</entry>
      <entry>Owner of the database, initially who created it</entry>
     </row>

     <row>
      <entry>encoding</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>Character/multibyte encoding for this database</entry>
     </row>

     <row>
      <entry>datistemplate</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       If true then this database can be used in the
       <quote>TEMPLATE</quote> clause of <command>CREATE
       DATABASE</command> to create the new database as a clone of
       this one.
      </entry>
     </row>

     <row>
      <entry>datallowconn</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       If false then no one can connect to this database.  This is
       used to protect the template0 database from being altered.
      </entry>
     </row>

     <row>
      <entry>datlastsysoid</entry>
      <entry><type>oid</type></entry>
      <entry></entry>
      <entry>
       Last oid in existence after the database was created; useful
       particularly to <application>pg_dump</application>
      </entry>
     </row>

     <row>
      <entry>datpath</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>
       If the database is stored at an alternative location then this
       records the location.  It's either an environment variable name
       or an absolute path, depending how it was entered.
      </entry>
     </row>
    </tbody>
   </tgroup>
  </table>
 </section>


 <section id="catalog-pg-description">
  <title>pg_description</title>

  <para>
   The pg_description table can store an optional description or
   comment for each database object.  Descriptions can be manipulated
   with the <command>COMMENT</command> command.  Client applications
   can view the descriptions by joining with this table.  Many builtin
   system objects have comments associated with them that are shown by
   <application>psql</application>'s <literal>\d</literal> commands.
  </para>

  <table>
   <title>pg_description Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>objoid</entry>
      <entry><type>oid</type></entry>
      <entry>any oid attribute</entry>
      <entry>The oid of the object this description pertains to</entry>
     </row>

     <row>
      <entry>description</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>Arbitrary text that serves as the description of this object.</entry>
     </row>
    </tbody>
   </tgroup>
  </table>

 </section>


 <section id="catalog-pg-group">
  <title>pg_group</title>

  <para>
   This catalog defines groups and stores what users belong to what
   groups.  Groups are created with the <command>CREATE
   GROUP</command> command.  Consult the <citetitle>Administrator's
   Guide</citetitle> for information about user permission management.
  </para>

  <table>
   <title>pg_group Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>groname</entry>
      <entry><type>name</type></entry>
      <entry></entry>
      <entry>Name of the group</entry>
     </row>

     <row>
      <entry>grosysid</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>An arbitrary number to identify this group</entry>
     </row>

     <row>
      <entry>grolist</entry>
      <entry><type>int4[]</type></entry>
      <entry>pg_shadow.usesysid</entry>
      <entry>An array containing the ids of the users in this group</entry>
     </row>
    </tbody>
   </tgroup>
  </table>

 </section>


 <section id="catalog-pg-index">
  <title>pg_index</title>

  <para>
   <structname>pg_index</structname> contains part of the information
   about indexes.  The rest is mostly in
   <structname>pg_class</structname>.
  </para>

  <table>
   <title>pg_index Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>indexrelid</entry>
      <entry><type>oid</type></entry>
      <entry>pg_class.oid</entry>
      <entry>The oid of the pg_class entry for this index</entry>
     </row>

     <row>
      <entry>indrelid</entry>
      <entry><type>oid</type></entry>
      <entry>pg_class.oid</entry>
      <entry>The oid of the pg_class entry for the table this index is for</entry>
     </row>

     <row>
      <entry>indproc</entry>
      <entry><type>oid</type></entry>
      <entry>pg_proc.oid</entry>
      <entry>The registered procedure if this is a functional index</entry>
     </row>

     <row>
      <entry>indkey</entry>
      <entry><type>int2vector</type></entry>
      <entry>pg_attribute.attnum</entry>
      <entry>
       This is an vector (array) of up to
       <symbol>INDEX_MAX_KEYS</symbol> values that indicate which
       table columns this index pertains to.  For example a value of
       <literal>1 3</literal> would mean that the first and the third
       column make up the index key.
      </entry>
     </row>

     <row>
      <entry>indclass</entry>
      <entry><type>oidvector</type></entry>
      <entry>pg_opclass.oid</entry>
      <entry>
       For each column in the index key this contains a reference to
       the <quote>operator class</quote> to use.  See
       <structname>pg_opclass</structname> for details.
      </entry>
     </row>

     <row>
      <entry>indisclustered</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>unused</entry>
     </row>

     <row>
      <entry>indislossy</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>???</entry>
     </row>

     <row>
      <entry>indisunique</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>If true, this is a unique index.</entry>
     </row>

     <row>
      <entry>indisprimary</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>If true, this index is a unique index that represents the primary key of the table.</entry>
     </row>

     <row>
      <entry>indreference</entry>
      <entry><type>oid</type></entry>
      <entry></entry>
      <entry>unused</entry>
     </row>

     <row>
      <entry>indpred</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>Query plan for partial index predicate (not functional)</entry>
     </row>
    </tbody>
   </tgroup>
  </table>

 </section>


 <section id="catalog-pg-inherits">
  <title>pg_inherits</title>

  <para>
   This catalog records information about table inheritance hierarchies.
  </para>

  <table>
   <title>pg_inherits Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>inhrelid</entry>
      <entry><type>oid</type></entry>
      <entry>pg_class.oid</entry>
      <entry>
       This is the reference to the subtable, that is, it records the
       fact that the identified table is inherited from some other
       table.
      </entry>
     </row>

     <row>
      <entry>inhparent</entry>
      <entry><type>oid</type></entry>
      <entry>pg_class.oid</entry>
      <entry>
       This is the reference to the parent table, from which the table
       referenced by <structfield>inhrelid</structfield> inherited
       from.
      </entry>
     </row>

     <row>
      <entry>inhseqno</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>
       If there is more than one subtable/parent pair (multiple
       inheritance), this number tells the order in which the
       inherited columns are to be arranged.  The count starts at 1.
      </entry>
     </row>
    </tbody>
   </tgroup>
  </table>

 </section>


 <section id="catalog-pg-language">
  <title>pg_language</title>

  <para>
   <structname>pg_language</structname> registers call interfaces or
   languages in which you can write functions or stored procedures.
   See under <command>CREATE LANGUAGE</command> and in the
   <citetitle>Programmer's Guide</citetitle> for more information
   about language handlers.
  </para>

  <table>
   <title>pg_language Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>lanname</entry>
      <entry><type>name</type></entry>
      <entry></entry>
      <entry>Name of the language (to be specified when creating a function)</entry>
     </row>

     <row>
      <entry>lanispl</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       This is false for internal languages (such as SQL) and true for
       dynamically loaded language handler modules.  It essentially
       means that, if it is true, the language may be dropped.
      </entry>
     </row>

     <row>
      <entry>lanpltrusted</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       This is a trusted language.  See under <command>CREATE
       LANGUAGE</command> what this means.  If this is an internal
       language (<structfield>lanispl</structfield> is false) then
       this field is meaningless.
      </entry>
     </row>

     <row>
      <entry>lanplcallfoid</entry>
      <entry><type>oid</type></entry>
      <entry>pg_proc.oid</entry>
      <entry>
       For non-internal languages this references the language
       handler, which is a special function that is responsible for
       executing all functions that are written in the particular
       language.
      </entry>
     </row>

     <row>
      <entry>lancompiler</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>not useful</entry>
     </row>
    </tbody>
   </tgroup>
  </table>

 </section>


 <section id="catalog-pg-operator">
  <title>pg_operator</title>

  <para>
   See <command>CREATE OPERATOR</command> and the
   <citetitle>Programmer's Guide</citetitle> for details on these
   operator parameters.
  </para>

  <table>
   <title>pg_operator Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>oprname</entry>
      <entry><type>name</type></entry>
      <entry></entry>
      <entry>Name of the operator</entry>
     </row>

     <row>
      <entry>oprowner</entry>
      <entry><type>int4</type></entry>
      <entry>pg_shadow.usesysid</entry>
      <entry>Owner (creator) of the operator</entry>
     </row>

     <row>
      <entry>oprprec</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>unused</entry>
     </row>

     <row>
      <entry>oprkind</entry>
      <entry><type>char</type></entry>
      <entry></entry>
      <entry>
       'b' = infix (<quote>both</quote>), 'l' = prefix
       (<quote>left</quote>), 'r' = postfix (<quote>right</quote>)
      </entry>
     </row>

     <row>
      <entry>oprisleft</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>unused</entry>
     </row>

     <row>
      <entry>oprcanhash</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>This operator supports hash joins.</entry>
     </row>

     <row>
      <entry>oprleft</entry>
      <entry><type>oid</type></entry>
      <entry>pg_type.oid</entry>
      <entry>Type of the left operand</entry>
     </row>

     <row>
      <entry>oprright</entry>
      <entry><type>oid</type></entry>
      <entry>pg_type.oid</entry>
      <entry>Type of the right operand</entry>
     </row>

     <row>
      <entry>oprresult</entry>
      <entry><type>oid</type></entry>
      <entry>pg_type.oid</entry>
      <entry>Type of the result</entry>
     </row>

     <row>
      <entry>oprcom</entry>
      <entry><type>oid</type></entry>
      <entry>pg_operator.oid</entry>
      <entry>Commutator of this operator, if any</entry>
     </row>

     <row>
      <entry>oprnegate</entry>
      <entry><type>oid</type></entry>
      <entry>pg_operator.oid</entry>
      <entry>Negator of this operator, if any</entry>
     </row>

     <row>
      <entry>oprlsortop</entry>
      <entry><type>oid</type></entry>
      <entry>pg_operator.oid</entry>
      <entry>
       If this operator supports merge joins, the operator that sorts
       the type of the left-hand operand
      </entry>
     </row>

     <row>
      <entry>oprrsortop</entry>
      <entry><type>oid</type></entry>
      <entry>pg_operator.oid</entry>
      <entry>
       If this operator supports merge joins, the operator that sorts
       the type of the right-hand operand
      </entry>
     </row>

     <row>
      <entry>oprcode</entry>
      <entry><type>regproc</type></entry>
      <entry></entry>
      <entry>Function that implements this operator</entry>
     </row>

     <row>
      <entry>oprrest</entry>
      <entry><type>regproc</type></entry>
      <entry></entry>
      <entry>Restriction selectivity estimation function for this operator</entry>
     </row>

     <row>
      <entry>oprjoin</entry>
      <entry><type>regproc</type></entry>
      <entry></entry>
      <entry>Join selectivity estimation function for this operator</entry>
     </row>
    </tbody>
   </tgroup>
  </table>

 </section>


 <section id="catalog-pg-proc">
  <title>pg_proc</title>

  <para>
   This catalog stores information about functions (or procedures).
   The description of <command>CREATE FUNCTION</command> and the
   <citetitle>Programmer's Guide</citetitle> contain more information
   about the meaning of some fields.
  </para>

  <table>
   <title>pg_proc Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>proname</entry>
      <entry><type>name</type></entry>
      <entry></entry>
      <entry>Name of the function</entry>
     </row>

     <row>
      <entry>proowner</entry>
      <entry><type>int4</type></entry>
      <entry>pg_shadow.usesysid</entry>
      <entry>Owner (creator) of the function</entry>
     </row>

     <row>
      <entry>prolang</entry>
      <entry><type>oid</type></entry>
      <entry>pg_language.oid</entry>
      <entry>Implementation language or call interface of this function</entry>
     </row>

     <row>
      <entry>proisinh</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>unused</entry>
     </row>

     <row>
      <entry>proistrusted</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>not functional</entry>
     </row>

     <row>
      <entry>proiscachable</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>Function returns same result for same input values</entry>
     </row>

     <row>
      <entry>proisstrict</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       Function returns null if any call argument is null.  In that
       case the function won't actually be called at all.  Functions
       that are not <quote>strict</quote> must be prepared to handle
       null inputs.
      </entry>
     </row>

     <row>
      <entry>pronargs</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>Number of arguments</entry>
     </row>

     <row>
      <entry>proretset</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>Function returns a set (probably not functional)</entry>
     </row>

     <row>
      <entry>prorettype</entry>
      <entry><type>oid</type></entry>
      <entry>pg_type.oid</entry>
      <entry>Data type of the return value (0 if the function does not return a value)</entry>
     </row>

     <row>
      <entry>proargtypes</entry>
      <entry><type>oidvector</type></entry>
      <entry>pg_type.oid</entry>
      <entry>A vector with the data types of the function arguments</entry>
     </row>

     <row>
      <entry>probyte_pct</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>dead code</entry>
     </row>

     <row>
      <entry>properbyte_pct</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>dead code</entry>
     </row>

     <row>
      <entry>propercall_pct</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>dead code</entry>
     </row>

     <row>
      <entry>prooutin_ratio</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>dead code</entry>
     </row>

     <row>
      <entry>prosrc</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>
       This tells the function handler how to invoke the function.  It
       might be the actual source code of the function for interpreted
       languages, a link symbol, a file name, or just about anything
       else, depending the implementation language/call convention.
      </entry>
     </row>

     <row>
      <entry>probin</entry>
      <entry><type>bytea</type></entry>
      <entry></entry>
      <entry>?</entry>
     </row>
    </tbody>
   </tgroup>
  </table>

 </section>


 <section id="catalog-pg-relcheck">
  <title>pg_relcheck</title>

  <para>
   This system catalog stores CHECK constraints on tables.  (Column
   constraints are not treated specially.  Every column constraint is
   equivalent to some table constraint.)  See under <command>CREATE
   TABLE</command> for more information.
  </para>

  <table>
   <title>pg_relcheck Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>rcrelid</entry>
      <entry><type>oid</type></entry>
      <entry>pg_class.oid</entry>
      <entry>The table this check constraint is on</entry>
     </row>

     <row>
      <entry>rcname</entry>
      <entry><type>name</type></entry>
      <entry></entry>
      <entry>Constraint name</entry>
     </row>

     <row>
      <entry>rcbin</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>An internal representation of the constraint expression</entry>
     </row>

     <row>
      <entry>rcsrc</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>A human-readable representation of the consraint expression</entry>
     </row>
    </tbody>
   </tgroup>
  </table>

  <note>
   <para>
    <structname>pg_class</structname>.<structfield>relchecks</structfield>
    needs to match up with the entries in this table.
   </para>
  </note>

 </section>


 <section id="catalog-pg-shadow">
  <title>pg_shadow</title>

  <para>
   <structname>pg_shadow</structname> contains information about
   database users.  The name stems from the fact that this table
   should not be readable by the public since it contains passwords.
   <structname>pg_user</structname> is a view on
   <structname>pg_shadow</structname> that blanks out the password field.
  </para>

  <para>
   The <citetitle>Administrator's Guide</citetitle> contains detailed
   information about user and permission management.
  </para>

  <table>
   <title>pg_shadow Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>usename</entry>
      <entry><type>name</type></entry>
      <entry></entry>
      <entry>User name</entry>
     </row>

     <row>
      <entry>usesysid</entry>
      <entry><type>int4</type></entry>
      <entry></entry>
      <entry>User id (arbitrary number used to reference this user)</entry>
     </row>

     <row>
      <entry>usecreatedb</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>User may create databases</entry>
     </row>

     <row>
      <entry>usetrace</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>not used</entry>
     </row>

     <row>
      <entry>usesuper</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>User is a superuser</entry>
     </row>

     <row>
      <entry>usecatupd</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       User may update system catalogs.  (Even a superuser may not do
       this unless this attribute is true.)
      </entry>
     </row>

     <row>
      <entry>passwd</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>Password</entry>
     </row>

     <row>
      <entry>valuntil</entry>
      <entry><type>abstime</type></entry>
      <entry></entry>
      <entry>Account expiry time (only used for password authentication)</entry>
     </row>
    </tbody>
   </tgroup>
  </table>

 </section>


 <section id="catalog-pg-type">
  <title>pg_type</title>

  <table>
   <title>pg_type Columns</title>

   <tgroup cols=4>
    <thead>
     <row>
      <entry>Name</entry>
      <entry>Type</entry>
      <entry>References</entry>
      <entry>Description</entry>
     </row>
    </thead>

    <tbody>
     <row>
      <entry>typname</entry>
      <entry><type>name</type></entry>
      <entry></entry>
      <entry>Data type name</entry>
     </row>

     <row>
      <entry>typowner</entry>
      <entry><type>int4</type></entry>
      <entry>pg_shadow.usesysid</entry>
      <entry>Owner (creator) of the type</entry>
     </row>

     <row>
      <entry>typlen</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>Length of the storage representation of the type, -1 if variable length</entry>
     </row>

     <row>
      <entry>typprtlen</entry>
      <entry><type>int2</type></entry>
      <entry></entry>
      <entry>unused</entry>
     </row>

     <row>
      <entry>typbyval</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>
       <structfield>typbyval</structfield> determines whether internal
       routines pass a value of this type by value or by reference.
       Only <type>char</type>, <type>short</type>, and
       <type>int</type> equivalent items can be passed by value, so if
       the type is not 1, 2, or 4 bytes long, Postgres does not have
       the option of passing by value and so
       <structfield>typbyval</structfield> had better be false.
       Variable-length types are always passed by reference. Note that
       <structfield>typbyval</structfield> can be false even if the
       length would allow pass-by-value; this is currently true for
       type <type>float4</type>, for example.
      </entry>
     </row>

     <row>
      <entry>typtype</entry>
      <entry><type>char</type></entry>
      <entry></entry>
      <entry>
       <structfield>typtype</structfield> is <literal>b</literal> for
       a basic type and <literal>c</literal> for a catalog type (i.e.,
       a table).  If <structfield>typtype</structfield> is
       <literal>c</literal>, <structfield>typrelid</structfield> is
       the OID of the type's entry in
       <structname>pg_class</structname>.
      </entry>
     </row>

     <row>
      <entry>typisdefined</entry>
      <entry><type>bool</type></entry>
      <entry></entry>
      <entry>???</entry>
     </row>

     <row>
      <entry>typdelim</entry>
      <entry><type>char</type></entry>
      <entry></entry>
      <entry>Character that separates two values of this type when parsing array input</entry>
     </row>

     <row>
      <entry>typrelid</entry>
      <entry><type>oid</type></entry>
      <entry>pg_class.oid</entry>
      <entry>
       If this is a catalog type (see
       <structfield>typtype</structfield>), then this field points to
       the <structfield>pg_class</structfield> entry that defines the
       corresponding table. A table could theoretically be used as a
       composite data type, but this is not fully functional.
      </entry>
     </row>

     <row>
      <entry>typelem</entry>
      <entry><type>oid</type></entry>
      <entry>pg_type.oid</entry>
      <entry>
       If <structfield>typelem</structfield> is not 0 then it
       identifies another row in <structname>pg_type</structname>.
       The current type can then be subscripted like an array yielding
       values of type <structfield>typelem</structfield>. A non-zero
       <structfield>typelem</structfield> does not guarantee this type
       to be a <quote>real</quote> array type; some ordinary
       fixed-length types can also be subscripted (e.g.,
       <type>oidvector</type>). Variable-length types can
       <emphasis>not</emphasis> be turned into pseudo-arrays like
       that. Hence, the way to determine whether a type is a
       <quote>true</quote> array type is typelem != 0 and typlen < 0.
      </entry>
     </row>

     <row>
      <entry>typinput</entry>
      <entry><type>regproc</type></entry>
      <entry></entry>
      <entry>Input function</entry>
     </row>

     <row>
      <entry>typoutput</entry>
      <entry><type>regproc</type></entry>
      <entry></entry>
      <entry>Output function</entry>
     </row>

     <row>
      <entry>typreceive</entry>
      <entry><type>regproc</type></entry>
      <entry></entry>
      <entry>unused</entry>
     </row>

     <row>
      <entry>typsend</entry>
      <entry><type>regproc</type></entry>
      <entry></entry>
      <entry>unused</entry>
     </row>

     <row>
      <entry>typalign</entry>
      <entry><type>char</type></entry>
      <entry></entry>
      <entry><para>

       <structfield>typalign</structfield> is the alignment required
       when storing a value of this type.  It applies to storage on
       disk as well as most representations of the value inside
       Postgres.  When multiple values are stored consecutively, such
       as in the representation of a complete row on disk, padding is
       inserted before a datum of this type so that it begins on the
       specified boundary.  The alignment reference is the beginning
       of the first datum in the sequence.
      </para><para>
       Possible values are:
       <itemizedlist>
        <listitem>
         <para>'c' = CHAR alignment, i.e., no alignment needed.</para>
        </listitem>
        <listitem>
         <para>'s' = SHORT alignment (2 bytes on most machines).</para>
        </listitem>
        <listitem>
         <para>'i' = INT alignment (4 bytes on most machines).</para>
        </listitem>
        <listitem>
         <para>'d' = DOUBLE alignment (8 bytes on many machines, but by no means all).</para>
        </listitem>
       </itemizedlist>
      </para><note>
       <para>
        For types used in system tables, it is critical that the size
        and alignment defined in <structname>pg_type</structname>
        agree with the way that the compiler will lay out the field in
        a struct representing a table row.
       </para>
      </note></entry>
     </row>

     <row>
      <entry>typstorage</entry>
      <entry><type>char</type></entry>
      <entry></entry>
      <entry><para>
       <structfield>typstorage</structfield> tells for variable-length
       types (those with <structfield>typlen</structfield> = -1) if
       the type is prepared for toasting and what the default strategy
       for attributes of this type should be.
       Possible values are
       <itemizedlist>
        <listitem>
         <para>'p': Value must always be stored plain.</para>
        </listitem>
        <listitem>
         <para>
          'e': Value can be stored in a <quote>secondary</quote>
          relation (if relation has one, see
          <structname>pg_class</structname>.<structfield>reltoastrelid</structfield>).
         </para>
        </listitem>
        <listitem>
         <para>'m': Value can be stored compressed inline.</para>
        </listitem>
        <listitem>
         <para>'x': Value can be stored compressed inline or in <quote>secondary</quote>.</para>
        </listitem>
       </itemizedlist>
       Note that 'm' fields can also be moved out to secondary
       storage, but only as a last resort ('e' and 'x' fields are
       moved first).
      </para></entry>
     </row>

     <row>
      <entry>typdefault</entry>
      <entry><type>text</type></entry>
      <entry></entry>
      <entry>???</entry>
     </row>
    </tbody>
   </tgroup>
  </table>

 </section>

</chapter>