Data::Dmp(3) Dump Perl data structures as Perl code


This document describes version 0.20 of Data::Dmp (from Perl distribution Data-Dmp), released on 2016-06-03.


use Data::Dmp; # exports dd() and dmp()
dd [1, 2, 3]; # prints "[1,2,3]"
$a = dmp({a => 1}); # -> "{a=>1}"


Data::Dmp is a Perl dumper like Data::Dumper. It's compact (only about 175 lines of code long), starts fast and does not use any non-core modules except Regexp::Stringify when dumping regexes. It produces compact single-line output (similar to Data::Dumper::Concise). It roughly has the same speed as Data::Dumper (usually a bit faster for smaller structures) and faster than Data::Dump, but does not offer the various formatting options. It supports dumping objects, regexes, circular structures, coderefs. Its code is first based on Data::Dump: I removed all the parts that I don't need, particularly the pretty formatting stuffs) and added some features that I need like proper regex dumping and coderef deparsing.


$Data::Dmp::OPT_PERL_VERSION => str (default: 5.010)

Set target Perl version. If you set this to, say 5.010, then the dumped code will keep compatibility with Perl 5.10.0. This is used in the following ways:
  • passed to Regexp::Stringify
  • when dumping code references

    For example, in perls earlier than 5.016, does not understand:

     no feature ':all';

    so we replace it with:

     no feature;

$Data::Dmp::OPT_REMOVE_PRAGMAS => bool (default: 0)

If set to 1, then pragmas at the start of coderef dump will be removed. Coderef dump is produced by B::Deparse and is of the form like:

 sub { use feature 'current_sub', 'evalbytes', 'fc', 'say', 'state', 'switch', 'unicode_strings', 'unicode_eval'; $a <=> $b }

If you want to dump short coderefs, the pragmas might be distracting. You can turn turn on this option which will make the above dump become:

 sub { $a <=> $b }

Note that without the pragmas, the dump might be incorrect.


                      Rate    Data::Dump Data::Dumper Data::Dmp
 Data::Dump    21199+-23/s            --       -67.3%    -76.4%
 Data::Dumper 64920+-110/s 206.24+-0.62%           --    -27.7%
 Data::Dmp     89744+-27/s 323.34+-0.47% 38.24+-0.24%        --
                        Rate    Data::Dump   Data::Dmp Data::Dumper
 Data::Dump    2667.7+-2.7/s            --      -76.0%       -76.2%
 Data::Dmp    11131.4+-3.1/s 317.27+-0.43%          --        -0.6%
 Data::Dumper    11199+-57/s   319.8+-2.2% 0.61+-0.51%           --
 Some mixed structure:
                     Rate Data::Dump    Data::Dmp Data::Dumper
 Data::Dump    6096+-16/s         --       -68.4%       -80.5%
 Data::Dmp    19284+-39/s  216.3+-1%           --       -38.5%
 Data::Dumper 31335+-94/s    414+-2% 62.49+-0.58%           --


dd($data, ...) => $data ...

Exported by default. Like "Data::Dump"'s "dd" (a.k.a. "dump"), print one or more data to STDOUT. Unlike "Data::Dump"'s "dd", it always prints and return the original data (like XXX), making it convenient to insert into expressions. This also removes ambiguity and saves one "wantarray()" call.

dmp($data, ...) => $str

Exported by default. Return dump result as string. Unlike "Data::Dump"'s "dd" (a.k.a. "dump"), it never prints and only return the data.


When to use Data::Dmp? How does it compare to other dumper modules?

Data::Dmp might be suitable for you if you want a relatively fast pure-Perl data structure dumper to eval-able Perl code. It produces compact, single-line Perl code but offers little/no formatting options. Data::Dmp and Data::Dump module family usually produce Perl code that is ``more eval-able'', e.g. it can recreate circular structure.

Data::Dump produces visually nicer output (some alignment, use of range operator to shorten lists, use of base64 for binary data, etc) but no built-in option to produce compact/single-line output. It's more suitable for debugging. It's also relatively slow. I usually use its variant, Data::Dump::Color, for console debugging.

Data::Dumper is a core module, offers a lot of formatting options (like disabling hash key sorting, setting verboseness/indent level, and so on) but you usually have to configure it quite a bit before it does exactly like you want (that's why there are modules on CPAN that are just wrapping Data::Dumper with some configuration, like Data::Dumper::Concise et al). It does not support dumping Perl code that can recreate circular structures.

Of course, dumping to eval-able Perl code is slow (not to mention the cost of re-loading the code back to in-memory data, via eval-ing) compared to dumping to JSON, YAML, Sereal, or other format. So you need to decide first whether this is the appropriate route you want to take. (But note that there is also Data::Dumper::Limited and Data::Undump which uses a format similar to Data::Dumper but lets you load the serialized data without eval-ing them, thus achieving the speed comparable to JSON::XS).


Please visit the project's homepage at <>.


Source repository is at <>.


Please report any bugs or feature requests on the bugtracker website <>

When submitting a bug or request, please include a test-file or a patch to an existing test-file that illustrates the bug or desired feature.


perlancar <[email protected]>


This software is copyright (c) 2016 by [email protected].

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.