postgresql/doc/src/sgml/ecpg.sgml

<!--
$Header: /cvsroot/pgsql/doc/src/sgml/ecpg.sgml,v 1.16 2000/09/18 20:11:36 petere Exp $
-->

 <chapter>
  <docinfo>
   <authorgroup>
    <author>
     <firstname>Linus</firstname>
     <surname>Tolke</surname>
    </author>
    <author>
     <firstname>Michael</firstname>
     <surname>Meskes</surname>
    </author>
   </authorgroup>
   <copyright>
    <year>1996-1997</year>
    <holder>Linus Tolke</holder>
   </copyright>
   <copyright>
    <year>1998</year>
    <holder>Michael Meskes</holder>
   </copyright>
   <date>Transcribed 1998-02-12</date>
  </docinfo>

  <title><application>ecpg</application> - Embedded <acronym>SQL</acronym> 
   in <acronym>C</acronym></title>

  <para>
   This describes an embedded <acronym>SQL</acronym> in <acronym>C</acronym> 
   package for <productname>Postgres</productname>.
   It is written by <ulink url="mailto:linus@epact.se">Linus Tolke</ulink>
   and <ulink url="mailto:meskes@debian.org">Michael Meskes</ulink>.
   The package is installed with the <productname>Postgres</> distribution.

   <note>
    <para>
     Permission is granted to copy and use in the same way as you are allowed
     to copy and use the rest of <productname>PostgreSQL</productname>.
    </para>
   </note>
  </para>

  <sect1>
   <title>Why Embedded <acronym>SQL</acronym>?</title>

   <para>
    Embedded <acronym>SQL</acronym> has some small advantages over other ways 
    to handle <acronym>SQL</acronym>
    queries. It takes care of all the tedious moving of information to and
    from variables in your <acronym>C</acronym> program. 
    Many <acronym>RDBMS</acronym> packages
    support this embedded language.
   </para>

   <para> There is an ANSI-standard describing how the embedded language should
    work. <application>ecpg</application> was designed to meet this standard 
    as much as possible. So it is
    possible to port programs with embedded <acronym>SQL</acronym> written for
    other <acronym>RDBMS</acronym> packages to
    <productname>Postgres</productname> and thus promoting the spirit of free
    software.
   </para>
  </sect1>

  <sect1>
   <title>The Concept</title>

   <para>
    You write your program in <acronym>C</acronym> with some 
    special <acronym>SQL</acronym> things.
    For declaring variables that can be used in 
    <acronym>SQL</acronym> statements you need to
    put them in a special declare section.
    You use a special syntax for the <acronym>SQL</acronym> queries.
   </para>

   <para>
    Before compiling you run the file through 
    the embedded <acronym>SQL</acronym> <acronym>C</acronym>
    preprocessor and it converts the <acronym>SQL</acronym> statements you used 
    to function
    calls with the variables used as arguments. Both variables that are used
    as input to the <acronym>SQL</acronym> statements and variables that will 
    contain the
    result are passed.
   </para>

   <para>
    Then you compile and at link time you link with a special library that
    contains the functions used. These functions (actually it is mostly one
    single function) fetches the information from the arguments, performs
    the <acronym>SQL</acronym> query using the ordinary interface 
    (<filename>libpq</filename>) and puts back
    the result in the arguments dedicated for output.
   </para>

   <para>
    Then you run your program and when the control arrives to 
    the <acronym>SQL</acronym>
    statement the <acronym>SQL</acronym> statement is performed against 
    the database and you
    can continue with the result.
   </para>
  </sect1>

  <sect1>
   <title>How To Use <application>ecpg</application></title>

   <para>
    This section describes how to use the <application>ecpg</application> tool.
   </para>

   <sect2>
    <title>Preprocessor</title>

    <para>
     The preprocessor is called <application>ecpg</application>. 
     After installation it resides in
     the <productname>Postgres</productname> <filename>bin/</filename> directory. 
    </para>
   </sect2>
   <sect2>
    <title>Library</title>

    <para>
     The <application>ecpg</application> library is called 
     <filename>libecpg.a</filename> or
     <filename>libecpg.so</filename>. Additionally, the library
     uses the <filename>libpq</filename> library for communication to the 
     <productname>Postgres</productname> server so you will
     have to link your program with <parameter>-lecpg -lpq</parameter>.
    </para>

    <para>
     The library has some methods that are "hidden" but that could prove very
     useful sometime.

     <itemizedlist>
      <listitem>
       <para>
	<function>ECPGdebug(int <replaceable>on</replaceable>, FILE
	 *<replaceable>stream</replaceable>)</function>
	turns on debug logging if called with the first argument non-zero.
	Debug logging is done on <replaceable>stream</replaceable>.
	Most <acronym>SQL</acronym> statement logs its arguments and result.
       </para>

       <para>
	The most important one (<function>ECPGdo</function>) 
	that is called on almost all <acronym>SQL</acronym>
	statements logs both its expanded string,
	i.e. the string
	with all the input variables inserted, and the result from the
	<productname>Postgres</productname> server.
	This can be very useful when searching for errors
	in your <acronym>SQL</acronym> statements.
       </para>
      </listitem>

      <listitem>
       <para>
	<function>ECPGstatus()</function>
	This method returns TRUE if we are connected to a database and FALSE if not.
       </para>
      </listitem>
     </itemizedlist>
    </para>
   </sect2>

   <sect2>
    <title>Error handling</title>

    <para>
     To be able to detect errors from the <productname>Postgres</productname> 
     server you include a line like
     <programlisting>
      exec sql include sqlca;
     </programlisting>
     in the include section of your file. This will define a struct and a
     variable with the name <parameter>sqlca</parameter> as following:
     <programlisting>
struct sqlca
{
 char sqlcaid[8];
 long sqlabc;
 long sqlcode;
 struct
 {
  int sqlerrml;
  char sqlerrmc[70];
 } sqlerrm;
 char sqlerrp[8];
 long sqlerrd[6];
 /* 0: empty                                         */
 /* 1: OID of processed tuple if applicable          */
 /* 2: number of rows processed in an INSERT, UPDATE */
 /*    or DELETE statement                           */
 /* 3: empty                                         */
 /* 4: empty                                         */
 /* 5: empty                                         */
 char sqlwarn[8];
 /* 0: set to 'W' if at least one other is 'W'       */
 /* 1: if 'W' at least one character string	     */
 /*    value was truncated when it was               */
 /*    stored into a host variable.                  */
 /* 2: empty                                         */
 /* 3: empty                                         */
 /* 4: empty                                         */
 /* 5: empty                                         */
 /* 6: empty                                         */
 /* 7: empty                                         */
 char sqlext[8];
} sqlca;
     </programlisting>
    </para>

    <para>
     If an error occured in the last <acronym>SQL</acronym> statement 
     then <parameter>sqlca.sqlcode</parameter>
     will be non-zero. If <parameter>sqlca.sqlcode</parameter> is less that 0
     then this is
     some kind of serious error, like the database definition does not match
     the query given. If it is bigger than 0 then this is a normal error like
     the table did not contain the requested row.
    </para>

    <para>
     sqlca.sqlerrm.sqlerrmc will contain a string that describes the error.
     The string ends with the line number
     in the source file.
    </para>

    <para>
     List of errors that can occur:

     <variablelist>
      <varlistentry>
       <term>-12, Out of memory in line %d.</term>
       <listitem>
	<para>
	 Does not normally occur. This is a sign that your virtual memory is
	 exhausted.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-200, Unsupported type %s on line %d.</term>
       <listitem>
	<para>
	 Does not normally occur. This is a sign that the preprocessor has
	 generated something that the library does not know about. Perhaps you
	 are running incompatible versions of the preprocessor and the library.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-201, Too many arguments line %d.</term>
       <listitem>
	<para>
	 This means that <productname>Postgres</productname> has returned more
	 arguments than we have
	 matching variables. Perhaps you have forgotten a couple of the host
	 variables in the <command>INTO :var1,:var2</command>-list.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-202, Too few arguments line %d.</term>
       <listitem>
	<para>
	 This means that <productname>Postgres</productname> has returned fewer
	 arguments than we have
	 host variables. Perhaps you have too many host variables in the 
	 <command>INTO :var1,:var2</command>-list.
	</para> 
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-203, Too many matches line %d.</term>
       <listitem>
	<para>
	 This means that the query has returned several lines but the
	 variables specified are no arrays. The <command>SELECT</command> you made
	 probably was not unique.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-204, Not correctly formatted int type: %s line %d.</term>
       <listitem>
	<para>
	 This means that the host variable is of an <type>int</type> type and the field
	 in the <productname>Postgres</productname> database is of another type and 
	 contains a value that cannot be interpreted as an <type>int</type>. 
	 The library uses <function>strtol</function>
	 for this conversion.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-205, Not correctly formatted unsigned type: %s line %d.</term>
       <listitem>
	<para>
	 This means that the host variable is of an <type>unsigned int</type> type and
	 the field in the <productname>Postgres</productname> database is of another 
	 type and contains a
	 value that cannot be interpreted as an <type>unsigned int</type>. The library
	 uses <function>strtoul</function> for this conversion.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-206, Not correctly formatted floating point type: %s line %d.</term>
       <listitem>
	<para>
	 This means that the host variable is of a <type>float</type> type and
	 the field in the <productname>Postgres</productname> database is of another 
	 type and contains a
	 value that cannot be interpreted as an <type>float</type>. The library
	 uses <function>strtod</function> for this conversion.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-207, Unable to convert %s to bool on line %d.</term>
       <listitem>
	<para>
	 This means that the host variable is of a <type>bool</type> type and
	 the field in the <productname>Postgres</productname> database is neither 't'
	 nor 'f'.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-208, Empty query line %d.</term>
       <listitem>
	<para>
	 <productname>Postgres</productname> returned PGRES_EMPTY_QUERY, probably
	 because the query indeed was empty.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-220, No such connection %s in line %d.</term>
       <listitem>
	<para>
	 The program tries to access a connection that does not exist.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-221, Not connected in line %d.</term>
       <listitem>
	<para>
	 The program tries to access a connection that does exist but is not open.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-230, Invalid statement name %s in line %d.</term>
       <listitem>
	<para>
	 The statement you are trying to use has not been prepared.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-400, Postgres error: %s line %d.</term>
       <listitem>
	<para>
	 Some <productname>Postgres</productname> error. 
	 The message contains the error message from the
	 <productname>Postgres</productname> backend.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-401, Error in transaction processing line %d. </term>
       <listitem>
	<para>
	 <productname>Postgres</productname> signalled to us that we cannot start,
	 commit or rollback the transaction.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>-402, connect: could not open database %s.</term>
       <listitem>
	<para>
	 The connect to the database did not work.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>100, Data not found line %d.</term>
       <listitem>
	<para>
	 This is a "normal" error that tells you that what you are quering cannot
	 be found or we have gone through the cursor.
	</para>
       </listitem>
      </varlistentry>

     </variablelist>
    </para>
   </sect2>
  </sect1>

  <sect1>
   <title>Limitations</title>

   <abstract>
    <para>
     What will never be included and why or what cannot be done with this
     concept.
    </para>
   </abstract>

   <para>
    <variablelist>
     <varlistentry>
      <term>Oracle's single tasking possibility</term>
      <listitem>
       <para>
	Oracle version 7.0 on AIX 3 uses the OS-supported locks on the shared
	memory segments and allows the application designer to link an
	application in a so called single tasking way. Instead of starting one
	client process per application process both the database part and the
	application part is run in the same process. In later versions of Oracle
	this is no longer supported.
       </para>

       <para>
	This would require a total redesign of the
	<productname>Postgres</productname> access model and
	that effort can not justify the performance gained.
       </para>
      </listitem>
     </varlistentry>
    </variablelist>
   </para>
  </sect1>

  <sect1>
   <title>Porting From Other <acronym>RDBMS</acronym> Packages</title>

   <para>
    The design of <application>ecpg</application> follows SQL standard. So
    porting from a standard RDBMS should not be a problem. Unfortunately there
    is no such thing as a standard RDBMS. So <application>ecpg</application>
    also tries to understand syntax additions as long as they do not create
    conflicts with the standard. 
   </para>

   <para>
    The following list shows all the known incompatibilities. If you find one
    not listed please notify 
    <ulink url="meskes@debian.org">Michael Meskes</ulink>.
    Note, however, that we list only incompatibilities from
    a precompiler of another RDBMS to <application>ecpg</application> and not
    additional <application>ecpg</application> features that these RDBMS do not
    have.
   </para>

   <para>
    <variablelist>
     <varlistentry>
      <term>Syntax of FETCH command</term>
      <listitem>
       <para>
	The standard syntax of the FETCH command is:
       </para>
       <para>
	FETCH [direction] [amount] IN|FROM <replaceable>cursor name</replaceable>.
       </para>
       <para>
	<application>ORACLE</application>, however, does not use the keywords IN
	resp. FROM. This feature cannot be added since it would create parsing
	conflicts.
       </para>
      </listitem>
     </varlistentry>
    </variablelist>
   </para>
  </sect1>

  <sect1>
   <title>For the Developer</title>

   <para>
    This section is for those who want to develop the 
    <application>ecpg</application> interface. It
    describes how the things work. The ambition is to make this section
    contain things for those that want to have a look inside and the section
    on How to use it should be enough for all normal questions.

    So, read this before looking at the internals of the 
    <application>ecpg</application>. If
    you are not interested in how it really works, skip this section.
   </para>

   <sect2>
    <title>ToDo List</title>

    <para>
     This version the preprocessor has some flaws:

     <variablelist>
      <varlistentry>
       <term>Library functions</term>
       <listitem>
	<para>
	 to_date et al. do not exists. But then <productname>Postgres</productname>
	 has some good conversion routines itself. So you probably won't miss these.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>Structures ans unions</term>
       <listitem>
	<para>
	 Structures and unions have to be defined in the declare section.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>Missing statements</term>
       <listitem>
	<para>
	 The following statements are not implemented thus far:
	 <variablelist>
	  <varlistentry>
	   <term> exec sql allocate</term>
	   <listitem>
	    <para>
	    </para>
	   </listitem>
	  </varlistentry>
	  <varlistentry>
	   <term> exec sql deallocate</term>
	   <listitem>
	    <para>
	    </para>
	   </listitem>
	  </varlistentry>
	  <varlistentry>
	   <term> SQLSTATE</term>
	   <listitem>
	    <para>
	    </para>
	   </listitem>
	  </varlistentry>
	 </variablelist>
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>message 'no data found'</term>
       <listitem>
	<para>
	 The error message for "no data" in an exec sql insert select from statement
	 has to be 100.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>sqlwarn[6]</term>
       <listitem>
	<para>
	 sqlwarn[6] should be 'W' if the PRECISION or SCALE value specified in a SET
	 DESCRIPTOR statement will be ignored.
	</para>
       </listitem>
      </varlistentry>
     </variablelist>
    </para>
   </sect2>

   <sect2>
    <title>The Preprocessor</title>

    <para>
     The first four lines written to the output are constant additions by ecpg.
     These are two comments and two include lines necessary for the interface to the
     library.
    </para>

    <para>
     Then the preprocessor works in one pass only, reading the input file and
     writing to the output as it goes along. Normally it just echoes
     everything to the output without looking at it further.
    </para>

    <para>
     When it comes to an <command>EXEC SQL</command> statements it intervenes and
     changes them depending on what it is. 
     The <command>EXEC SQL</command> statement can be one of these:

     <variablelist>
      <varlistentry>
       <term>Declare sections</term>
       <listitem>
	<para>
	 Declare sections begins with
	 <programlisting>
exec sql begin declare section;
	 </programlisting>
	 and ends with
	 <programlisting>
exec sql end declare section;
	 </programlisting>
	 In the section only variable declarations are allowed. Every variable
	 declare within this section is also entered in a list of variables
	 indexed on their name together with the corresponding type.
	</para>

	<para>
	 In particular the definition of a structure or union also has to be listed
	 inside a declare section. Otherwise <application>ecpg</application> cannot
	 handle these types since it simply does not know the definition.
	</para>

	<para>
	 The declaration is echoed to the file to make the variable a normal
	 C-variable also.
	</para>

	<para>
	 The special types VARCHAR and VARCHAR2 are converted into a named struct
	 for every variable. A declaration like:
	 <programlisting>
VARCHAR var[180];
	 </programlisting>
	 is converted into
	 <programlisting>
struct varchar_var { int len; char arr[180]; } var;
	 </programlisting>
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>Include statements</term>
       <listitem>
	<para>
	 An include statement looks like:
	 <programlisting>
exec sql include filename;
	 </programlisting>
	 Note that this is NOT the same as
	 <programlisting>
#include &lt;filename.h&gt;
	 </programlisting>
	</para>

	<para>
	 Instead the file specified is parsed by <application>ecpg</application>
	 itself. So the contents of the specified file is included in the resulting C
	 code. This way you are able to specify EXEC SQL commands in an include file.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>Connect statement</term>
       <listitem>
	<para>
	 A connect statement looks like:
	 <programlisting>
exec sql connect to <replaceable>connection target</replaceable>;
	 </programlisting>
	 It creates a connection to the specified database.
	</para>

	<para>
	 The <replaceable>connection target</replaceable> can be specified in the
	 following ways:
	 <variablelist>
	  <varlistentry>
	   <term>dbname[@server][:port][as <replaceable>connection
	     name</replaceable>][user <replaceable>user name</replaceable>]</term>
	   <listitem><para></para></listitem>
	  </varlistentry>

	  <varlistentry>
	   <term>tcp:postgresql://server[:port][/dbname][as
	    <replaceable>connection name</replaceable>][user <replaceable>user name</replaceable>]</term>
	   <listitem><para></para></listitem>
	  </varlistentry>

	  <varlistentry>
	   <term>unix:postgresql://server[:port][/dbname][as
	    <replaceable>connection name</replaceable>][user <replaceable>user name</replaceable>]</term>
	   <listitem><para></para></listitem>
	  </varlistentry>

	  <varlistentry>
	   <term><replaceable>character variable</replaceable>[as
	    <replaceable>connection name</replaceable>][user <replaceable>user name</replaceable>]</term>
	   <listitem><para></para></listitem>
	  </varlistentry>

	  <varlistentry>
	   <term><replaceable>character string</replaceable>[as
	    <replaceable>connection name</replaceable>][<replaceable>user</replaceable>]</term>
	   <listitem><para></para></listitem>
	  </varlistentry>

	  <varlistentry>
	   <term>default</term>
	   <listitem><para></para></listitem>
	  </varlistentry>

	  <varlistentry>
	   <term>user</term>
	   <listitem><para></para></listitem>
	  </varlistentry>
	 </variablelist>
	</para>

	<para>
	 There are also different ways to specify the user name:
	 <variablelist>
	  <varlistentry>
	   <term><replaceable>userid</replaceable></term>
	   <listitem><para></para></listitem>
	  </varlistentry>
	  <varlistentry>
	   <term><replaceable>userid</replaceable>/<replaceable>password</replaceable></term>
	   <listitem><para></para></listitem>
	  </varlistentry>
	  <varlistentry>
	   <term><replaceable>userid</replaceable> identified by <replaceable>password</replaceable></term>
	   <listitem><para></para></listitem>
	  </varlistentry>
	  <varlistentry>
	   <term><replaceable>userid</replaceable> using <replaceable>password</replaceable></term>
	   <listitem><para></para></listitem>
	  </varlistentry>
	 </variablelist>
	</para>

	<para> Finally the userid and the password. Each may be a constant text, a
	 character variable or a chararcter string.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>Disconnect statements</term>
       <listitem>
	<para>
	 A disconnect statement looks loke:
	 <programlisting>
exec sql disconnect [<replaceable>connection target</replaceable>];
	 </programlisting>
	 It closes the connection to the specified database.
	</para>

	<para>
	 The <replaceable>connection target</replaceable> can be specified in the
	 following ways:
	 <variablelist>
	  <varlistentry>
	   <term><replaceable>connection name</replaceable></term>
	   <listitem><para></para></listitem>
	  </varlistentry>
	  <varlistentry>
	   <term>default</term>
	   <listitem><para></para></listitem>
	  </varlistentry>
	  <varlistentry>
	   <term>current</term>
	   <listitem><para></para></listitem>
	  </varlistentry>
	  <varlistentry>
	   <term>all</term>
	   <listitem><para></para></listitem>
	  </varlistentry>
	 </variablelist>
	</para>
       </listitem>
      </varlistentry>

      <!--WARNING: FROM HERE ON THE TEXT IS OUTDATED!-->
      <varlistentry>
       <term>Open cursor statement</term>
       <listitem>
	<para>
	 An open cursor statement looks like:
	 <programlisting>
exec sql open <replaceable>cursor</replaceable>;
	 </programlisting>
	 and is ignore and not copied from the output.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>Commit statement</term>
       <listitem>
	<para>
	 A commit statement looks like
	 <programlisting>
exec sql commit;
	 </programlisting>
	 and is translated on the output to
	 <programlisting>
ECPGcommit(__LINE__);
	 </programlisting>
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>Rollback statement</term>
       <listitem>
	<para>
	 A rollback statement looks like
	 <programlisting>
exec sql rollback;
	 </programlisting>
	 and is translated on the output to
	 <programlisting>
ECPGrollback(__LINE__);
	 </programlisting>
	</para>
       </listitem>
      </varlistentry>

      <!--STARTING HERE IT IS OKAY AGAIN!-->
      <varlistentry>
       <term>Other statements</term>
       <listitem>
	<para>
	 Other <acronym>SQL</acronym> statements are other statements that start with 
	 <command>exec sql</command> and ends with <command>;</command>. 
	 Everything inbetween is treated
	 as an <acronym>SQL</acronym> statement and parsed for variable substitution.
	</para>

	<para>
	 Variable substitution occur when a symbol starts with a colon
	 (<command>:</command>). Then a variable with that name is looked for among
	 the variables that were previously declared within a declare section and
	 depending on the variable being for input or output the pointers to the
	 variables are written to the output to allow for access by the function.
	</para>

	<para>
	 For every variable that is part of the <acronym>SQL</acronym> request 
	 the function gets another ten arguments:

	 <simplelist>
	  <member>The type as a special symbol.</member>
	  <member>A pointer to the value or a pointer to the pointer.</member>
	  <member>The size of the variable if it is a char or varchar.</member>
	  <member>Number of elements in the array (for array fetches).</member>
	  <member>The offset to the next element in the array (for array fetches)</member>
	  <member>The type of the indicator variable as a special symbol.</member>
	  <member>A pointer to the value of the indicator variable or a pointer to the pointer of the indicator variable.</member>
	  <member>0.</member>
	  <member>Number of elements in the indicator array (for array fetches).</member>
	  <member>The offset to the next element in the indicator array (for array fetches)</member>
	 </simplelist>
	</para>

       </listitem>
      </varlistentry>
     </variablelist>
    </para>
   </sect2>

   <sect2>
    <title>A Complete Example</title>

    <para>
     Here is a complete example describing the output of the preprocessor of a
     file foo.pgc:
     <programlisting>
exec sql begin declare section;
int index;
int result;
exec sql end declare section;
...
exec sql select res into :result from mytable where index = :index;
     </programlisting>
     is translated into:
     <programlisting>
/* Processed by ecpg (2.6.0) */
/* These two include files are added by the preprocessor */
#include &lt;ecpgtype.h&gt;;
#include &lt;ecpglib.h&gt;;

/* exec sql begin declare section */

#line 1 "foo.pgc"

 int index;
 int result;
/* exec sql end declare section */
...
ECPGdo(__LINE__, NULL, "select  res  from mytable where index = ?     ",
	ECPGt_int,&amp;(index),1L,1L,sizeof(int),
        ECPGt_NO_INDICATOR, NULL , 0L, 0L, 0L, ECPGt_EOIT,
        ECPGt_int,&amp;(result),1L,1L,sizeof(int),
        ECPGt_NO_INDICATOR, NULL , 0L, 0L, 0L, ECPGt_EORT);
#line 147 "foo.pgc"

     </programlisting>
     (the indentation in this manual is added for readability and not
     something that the preprocessor can do.)
    </para>
   </sect2>

   <sect2>
    <title>The Library</title>

    <para>
     The most important function in the library is the <function>ECPGdo</function>
     function. It takes a variable amount of arguments. Hopefully we will not run
     into machines with limits on the amount of variables that can be
     accepted by a vararg function. This could easily add up to 50 or so
     arguments.
    </para>

    <para>
     The arguments are:

     <variablelist>
      <varlistentry>
       <term>A line number</term>
       <listitem>
	<para>
	 This is a line number for the original line used in error messages only.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>A string</term>
       <listitem>
	<para>
	 This is the <acronym>SQL</acronym> request that is to be issued. 
	 This request is modified
	 by the input variables, i.e. the variables that where not known at
	 compile time but are to be entered in the request. Where the variables
	 should go the string contains ";".
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>Input variables</term>
       <listitem>
	<para>
	 As described in the section about the preprocessor every input variable
	 gets ten arguments.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>ECPGt_EOIT</term>
       <listitem>
	<para>
	 An enum telling that there are no more input variables.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>Output variables</term>
       <listitem>
	<para>
	 As described in the section about the preprocessor every input variable
	 gets ten arguments. These variables are filled by the function.
	</para>
       </listitem>
      </varlistentry>

      <varlistentry>
       <term>ECPGt_EORT</term>
       <listitem>
	<para>
	 An enum telling that there are no more variables.
	</para>
       </listitem>
      </varlistentry>
     </variablelist>
    </para>

    <para>
     All the <acronym>SQL</acronym> statements are performed in one transaction
     unless you issue a commit transaction. To get this auto-transaction going
     the first statement or the first after statement after a commit or rollback
     always begins a transaction. To disable this feature per default use the
     <option>-t</option> option on the commandline.
    </para>

    <para>
     To be completed: entries describing the other entries.
    </para>
   </sect2>
  </sect1>
 </chapter>

<!-- Keep this comment at the end of the file
Local variables:
mode:sgml
sgml-omittag:nil
sgml-shorttag:t
sgml-minimize-attributes:nil
sgml-always-quote-attributes:t
sgml-indent-step:1
sgml-indent-data:t
sgml-parent-document:nil
sgml-default-dtd-file:"./reference.ced"
sgml-exposed-tags:nil
sgml-local-catalogs:("/usr/lib/sgml/catalog")
sgml-local-ecat-files:nil
End:
-->