View on
MetaCPAN is shutting down
For details read Perl NOC. After June 25th this page will redirect to
Max Maischein > DBIx-DataAudit-0.12 > DBIx::DataAudit



Annotate this POD

View/Report Bugs
Module Version: 0.12   Source   Latest Release: DBIx-DataAudit-0.13


DBIx::DataAudit - summarize column data for a table


  use DBIx::DataAudit;

  warn "Running audit for table $table";
  my $audit = DBIx::DataAudit->audit( dsn => 'dbi:SQLite:dbname=test.sqlite', table => 'test' );
  print $audit->as_text;

  # or
  print $audit->as_html;

This module provides a summary about the data contained in a table. It provides the descriptive statistics for every column. It's surprising how much bad data you find by looking at the minimum and maximum values of a column alone.

It tries to get the information in one table scan.


The module works by constructing an SQL statement that collects the information about the columns in a single full table scan.


You can specify which information is collected about every column by specifying the traits. The hierarchy of traits is as follows:

  any < ordered < numeric 
                < string

The following traits are collected for every column by default:

For columns that are recognized as ordered, the following additional traits are collected:

For columns that are recognized as numeric, the following additional traits are collected:

For columns that are recognized as string, the following additional traits are collected:


To customize some default behaviour, the some global variables are defined. Read the source to find their names.


The class implements the following methods:

__PACKAGE__->audit ARGS

Performs the data audit. Valid arguments are:

$audit->as_text RESULTS

Returns a table drawn as text with the results.

$audit->as_html RESULTS, TEMPLATE

Returns a HTML page with the results.

You can pass in a custom resultset or undef if you want the module to determine the results.

You can pass in a custom (Template) template if you want fancier rendering.


Returns a hash with the following three keys, suitable for using with whatever templating system you have:


Actually runs the SQL in the database.

$audit->column_type COLUMN

Returns the type for the column. The four valid types are any, ordered, numeric and string.

$audit->get_columns TABLE

Returns the names of the columns for the table TABLE. By default, the value of TABLE will be taken from the value passed to the constructor audit.

$audit->collect_column_info TABLE

Collects the information about the columns for the table TABLE from the DBI. By default, TABLE will be taken from the value passed to the constructor audit.

If your database driver does not implement the ->column_info method you are out of luck. A fatal error is raised by this method if ->column_info does not return anything.

For SQLite, DBD::SQLite::Amalgamation v3.6.1.2 includes the patch from Fey::Loader::SQLite, so if you want to use DBIx::DataAudit with SQLite, consider upgrading to DBD::SQLite::Amalgamation.

This method will raise warnings if it encounters a data type that it doesn't know yet. You can either patch the global variable %sql_type_map to add the type or submit a patch to me to add the type and its interpretation.

$audit->get_sql TABLE

Creates the SQL statement to collect the information. The default value for TABLE will be the table passed to the constructor audit.

If you encounter errors from your SQL engine, you may want to print the result of this method out.

$audit->trait_applies TRAIT, COLUMN

Checks whether a trait applies to a column.

A trait applies to a column if the trait type is any or if it is the same type as the column type as returned by get_column_type.

The method will raise an error if it is passed an unknown trait name. See the source code for how to add custom traits.


You can use this mail from the command line if you need a quick check of data:

  perl -MDBIx::DataAudit=dbi:SQLite:dbname=some/db.sqlite my_table [traits]

This could also incredibly useful if you want a breakdown of a csv-file:

  perl -MDBIx::DataAudit=dbi:AnyData:dbname=some/db.sqlite my_table [traits]

Unfortunately, that does not work yet, as I haven't found a convenient oneliner way to make a CSV file appear as database.



Max Maischein


Copyright 2008-2009 by Max Maischein


This module is released under the same terms as Perl itself.

syntax highlighting: