NAME

SQL::Translator::Producer::TT::Table -
Produces output using the Template Toolkit from a SQL schema, per table.

SYNOPSIS

  # Normal STDOUT version
  #
  my $translator     = SQL::Translator->new(
      from           => 'MySQL',
      filename       => 'foo_schema.sql',
      to             => 'TT::Table',
      producer_args  => {
          tt_table     => 'foo_table.tt',
      },
  );
  print $translator->translate;
  # To generate a file per table
  #
  my $translator     = SQL::Translator->new(
      from           => 'MySQL',
      filename       => 'foo_schema.sql',
      to             => 'TT::Table',
      producer_args  => {
          tt_table       => 'foo_table.tt.html',
          mk_files      => 1,
          mk_files_base => "./doc/tables",
          mk_file_ext   => ".html",
          on_exists     => "replace",
      },
  );
  #
  # ./doc/tables/ now contains the templated tables as $tablename.html
  #

DESCRIPTION

Produces schema output using a given Template Tookit template, processing that template for each table in the schema. Optionally allows you to write the result for each table to a separate file.

It needs one additional producer_arg of "tt_table" which is the file name of the template to use. This template will be passed a template var of "table", which is the current SQL::Translator::Schema::Table table we are producing, which you can then use to walk the schema via the methods documented in that module. You also get "schema" as a shortcut to the SQL::Translator::Schema for the table and "translator", the SQL::Translator object for this parse in case you want to get access to any of the options etc set here.

Here's a brief example of what the template could look like:

  [% table.name %]
  ================
  [% FOREACH field = table.get_fields %]
      [% field.name %]   [% field.data_type %]([% field.size %])
  [% END -%]

See t/data/template/table.tt for a more complete example.

You can also set any of the options used to initialize the Template object by adding them to your producer_args. See Template Toolkit docs for details of the options.

  $translator          = SQL::Translator->new(
      to               => 'TT',
      producer_args    => {
          ttfile       => 'foo_template.tt',
          INCLUDE_PATH => '/foo/templates/tt',
          INTERPOLATE  => 1,
      },
  );

If you set "mk_files" and its additional options the producer will write a separate file for each table in the schema. This is useful for producing things like HTML documentation where every table gets its own page (you could also use TTSchema producer to add an index page). It's also particularly good for code generation where you want to produce a class file per table.

OPTIONS

tt_table

File name of the template to run for each table.

mk_files

Set to true to output a file for each table in the schema (as well as returning the whole lot back to the Translalor and hence STDOUT). The file will be named after the table, with the optional "mk_files_ext" added and placed in the directory "mk_files_base".

mk_files_ext

Extension (without the dot) to add to the filename when using mk_files.

mk_files_base = DIR

Dir to build the table files into when using mk_files. Defaults to the current directory.

mk_file_dir

Set true and if the file needs to written to a directory that doesn't exist, it will be created first.

on_exists [Default:replace]

What to do if we are running with mk_files and a file already exists where we want to write our output. One of "skip", "die", "replace", "insert". The default is die.

replace - Over-write the existing file with the new one, clobbering anything already there.

skip - Leave the original file as it was and don't write the new version anywhere.

die - Die with an existing file error.

insert - Insert the generated output into the file between a set of special comments (defined by the following options.) Any code between the comments will be overwritten (ie the results from a previous produce) but the rest of the file is left alone (your custom code). This is particularly useful for code generation as it allows you to generate schema derived code and then add your own custom code to the file. Then when the schema changes you just re-produce to insert the new code.

insert_comment_start

The comment to look for in the file when on_exists is "insert". Default is "SQLF INSERT START". Must appear on it own line, with only whitespace either side, to be recognised.

insert_comment_end

The end comment to look for in the file when on_exists is "insert". Default is "SQLF INSERT END". Must appear on it own line, with only whitespace either side, to be recognised.

AUTHOR

Mark Addison <grommit@users.sourceforge.net>.

TODO

- Some tests for the various on exists options (they have been tested implicitly through use in a project but need some proper tests).

- More docs on code generation strategies.

- Better hooks for filename generation.