<application>psql</application>, are not prepared to
   handle binary cursors and expect data to come back in the text
   format.
  </para>

  <note>
   <para>
    When the client application uses the <quote>extended query</quote> protocol
    to issue a <command>FETCH</command> command, the Bind protocol message
    specifies whether data is to be retrieved in text or binary format.
    This choice overrides the way that the cursor is defined.  The concept
    of a binary cursor as such is thus obsolete when using extended query
    protocol &mdash; any cursor can be treated as either text or binary.
   </para>
  </note>

   <para>
    Unless <literal>WITH HOLD</literal> is specified, the cursor
    created by this command can only be used within the current
    transaction.  Thus, <command>DECLARE</command> without <literal>WITH
    HOLD</literal> is useless outside a transaction block: the cursor would
    survive only to the completion of the statement.  Therefore
    <productname>PostgreSQL</productname> reports an error if such a
    command is used outside a transaction block.
    Use
    <link linkend="sql-begin"><command>BEGIN</command></link> and
    <link linkend="sql-commit"><command>COMMIT</command></link>
    (or <link linkend="sql-rollback"><command>ROLLBACK</command></link>)
    to define a transaction block.
   </para>

   <para>
    If <literal>WITH HOLD</literal> is specified and the transaction
    that created the cursor successfully commits, the cursor can
    continue to be accessed by subsequent transactions in the same
    session.  (But if the creating transaction is aborted, the cursor
    is removed.)  A cursor created with <literal>WITH HOLD</literal>
    is closed when an explicit <command>CLOSE</command> command is
    issued on it, or the session ends.  In the current implementation,
    the rows represented by a held cursor are copied into a temporary
    file or memory area so that they remain available for subsequent
    transactions.
   </para>

   <para>
    <literal>WITH HOLD</literal> may not be specified when the query
    includes <literal>FOR UPDATE</literal> or <literal>FOR SHARE</literal>.
   </para>

   <para>
    The <literal>SCROLL</literal> option should be specified when defining a
    cursor that will be used to fetch backwards.  This is required by
    the SQL standard.  However, for compatibility with earlier
    versions, <productname>PostgreSQL</productname> will allow
    backward fetches without <literal>SCROLL</literal>, if the cursor's query
    plan is simple enough that no extra overhead is needed to support
    it. However, application developers are advised not to rely on
    using backward fetches from a cursor that has not been created
    with <literal>SCROLL</literal>.  If <literal>NO SCROLL</literal> is
    specified, then backward fetches are disallowed in any case.
   </para>

   <para>
    Backward fetches are also disallowed when the query
    includes <literal>FOR UPDATE</literal> or <literal>FOR SHARE</literal>; therefore
    <literal>SCROLL</literal> may not be specified in this case.
   </para>

   <caution>
    <para>
     Scrollable cursors may give unexpected
     results if they invoke any volatile functions (see <xref
     linkend="xfunc-volatility"/>).  When a previously fetched row is
     re-fetched, the functions might be re-executed, perhaps leading to
     results different from the first time.  It's best to
     specify <literal>NO SCROLL</literal> for a query involving volatile
     functions.  If that is not practical, one workaround
     is to declare the cursor <literal>SCROLL WITH HOLD</literal> and commit the
     transaction before reading any rows from it.  This will force the
     entire output of the cursor to be materialized in temporary storage,
     so that volatile functions are executed exactly once for each row.
    </para>
   </caution>

   <para>
    If the cursor's query includes <literal>FOR UPDATE</literal> or <literal>FOR