++ed by:

1 non-PAUSE user.

Author image Hugo Cornelis


Data::Transformator - transform nested Perl data structures.


   From one of the test cases:

    use Data::Transformator;

    my $tree
        = {
           e => [
                  e1 => {
                  e2 => {
                  e3 => {

    my $expected_data
        = {
           e => [
                  e1 => {

    my $transformation
        = Data::Transformator->new
             name => 'test_transform3',
             contents => $tree,
             apply_identity_transformation => {
                                               e => [

    my $transformed_data = $transformation->transform();

    use Data::Comparator qw(data_comparator);

    my $differences = data_comparator($transformed_data, $expected_data);

    if ($differences->is_empty())
        print "$0: 3: success\n";

        ok(1, '3: success');
        print "$0: 3: failed\n";

        ok(0, '3: failed');


Data::Transformator allows to transform a nested perl data structure -- the source data -- in a new nested perl data structure -- the result data. The source nested data structure can contain perl hashes, arrays and scalars, but should not be self-referential (if I remember well, the existing protection against self-referential data structures in the transformator engine is currently broken).


Running a transformation can be done in two essentially different ways:

  • extract data from an existing source : the transformation applies an identity transformation (meaning copy the source structure to the result), and during the process it uses selectors to select what data from the source to copy to the result. The result automatically inherits the structure from the source. This is called a selective transformation. Use the key 'apply_identity_transformation' to enable this mode.

  • copy data from source to result : the transformation must be told what the exact structure of the result is, before the result data can be inserted into the result. This is called a constructive transformation.

It is possible to combine the two above in a single run, but that is currently not tested enough to be sure it works alright, so be very careful with that.

To use Data::Transformator, you have to

  1. Construct the transformation with appropriate options:

    • Give the transformation a name, the name describes the purpose and/or activities for the transformation.

    • Tell the transformation if you want it to run as a selective transformation or not (option name 'apply_identity_transformation'). Transformations can always be used as constructive transformations.

    • Tell the transformation how to find the data source.

      • Or the data source is literal content (option name 'contents').

      • Or the data source is an object that implements a '->generate()' method (option name 'source'). Transformations implement themselves a '->generate()' method such that transformations can be cascaded easily.

    • Tell the transformation what to transform (selection) and how to transform (generate result). Data::Transformator uses code references to generate results, of alternatively simpler things as explained below. Whenever a code reference is used, it is called with the arguments (self, context, current_content). Here, self is the Data::Transformator object, context is an object that describes the current context in the source data structure, current_content is the generated result so far. $context->{path} contains a string with the path to the current element. Compononents of the path are '/' separated (unless overwritten with the constructor key 'separator'). This can be used for regular matching, and is especially handy using 'simple_transformators', see below.

      Following keys are available to the constructor of Data::Transformator:


      Is an array of simple_transformators. Each simple_transformator is a hash with a 'matcher' key that contains a regular expression that is matched with the path of the currently selected element. If there is a match, the selected subtree is put in the end result, under the value of the 'key' element of the simple_transformator (creating a hash in the result if necessary). If there is no 'key' element in the simple_transformator, a 'code' element is looked for, which is a code reference. The code is called to insert an appropriate result.


      Is an array of transformators. Each transformator is code reference that gets called as usual.


      Contains a nested perl data structure that reflects the structure of the source data. All data of the source that is selected by scalars that evaluate to true in the content of apply_identity_transformation, is inserted in the result. This key is very handy for selecting a set of small portions of data, if the structure of the source is known beforehand.

      Take a look at existing examples, e.g. in the unit tests of the transformation engine.

  2. Call the '->transform()' method on the transformator.

  3. Use the result data that is returned by the '->transform()' method.

The transformation library

There is a small transformation library embedded in the Data::Transformator. This library currently allows to

  • transform an array to be found somewhere in the data source to a hash in the result set.

  • transform a hash to be found somewhere in the data source to an array in the result set.

The library generates closures that work on the source data.


For the interested reader, please follow these reasoning steps:

  1. A database query of a relational database (using tables and nested tabled) can always be written out in one of the XML query dialects.

  2. Following from point 1: a database query can be expressed as a structured query.

  3. A structured query can be summarized as

    1. a selection of data from a preexisting data source.

    2. a structural simplification of the selection.

  4. Combining the above: a query can be defined as applying (1) a selective transformation and (2) a constructive transformation in sequence to a preexisting data source. This is called cascaded transformations.


Does only work with scalars, hashes and arrays. Support for self-referential structures seems broken at the moment.


Hugo Cornelis, hugo.cornelis@gmail.com

Copyright 2007 Hugo Cornelis.

This module is free software; you can redistribute it and/or modify it under the same terms as Perl itself.


Data::Merger(3), Data::Comparator(3), Clone(3)