YAML::PP - YAML 1.2 processor


WARNING: This is not yet stable.

Here are a few examples of the basic load and dump methods:

    use YAML::PP;
    my $ypp = YAML::PP->new;
    my $yaml = <<'EOM';
    --- # Document one is a mapping
    name: Tina
    age: 29
    favourite language: Perl

    --- # Document two is a sequence
    - plain string
    - 'in single quotes'
    - "in double quotes we have escapes! like \t and \n"
    - | # a literal block scalar
    - > # a folded block scalar
      this is all one
      single line because the
      linebreaks will be folded

    my @documents = $ypp->load_string($yaml);
    my @documents = $ypp->load_file($filename);

    my $yaml = $ypp->dump_string($data1, $data2);
    $ypp->dump_file($filename, $data1, $data2);

    # The loader offers JSON::PP::Boolean, or
    # perl 1/'' (currently default) for booleans
    my $ypp = YAML::PP->new(boolean => 'JSON::PP');
    my $ypp = YAML::PP->new(boolean => 'boolean');
    my $ypp = YAML::PP->new(boolean => 'perl');

    # Legacy interface
    use YAML::PP qw/ Load Dump LoadFile DumpFile /;
    my @documents = Load($yaml);
    my @documents = LoadFile($filename);
    my @documents = LoadFile($filehandle);
    my $yaml = = Dump(@documents);
    DumpFile($filename, @documents);
    DumpFile($filenhandle @documents);

    my $ypp = YAML::PP->new(schema => [qw/ JSON Perl /]);
    my $yaml = $yp->dump_string($data_with_perl_objects);

Some utility scripts, mostly useful for debugging:

    # Load YAML into a data structure and dump with Data::Dumper
    yamlpp5-load < file.yaml

    # Load and Dump
    yamlpp5-load-dump < file.yaml

    # Print the events from the parser in yaml-test-suite format
    yamlpp5-events < file.yaml

    # Parse and emit events directly without loading
    yamlpp5-parse-emit < file.yaml

    # Create ANSI colored YAML. Can also be useful for invalid YAML, showing
    # you the exact location of the error
    yamlpp5-highlight < file.yaml


YAML::PP is a modern, modular YAML processor.

It aims to support YAML 1.2 and YAML 1.1. See

YAML is a serialization language. The YAML input is called "YAML Stream". A stream consists of one or more "Documents", seperated by a line with a document start marker ---. A document optionally ends with the document end marker ....

This allows to process continuous streams additionally to a fixed input file or string.

The YAML::PP frontend will currently load all documents, and return only the last if called with scalar context.

The YAML backend is implemented in a modular way that allows to add custom handling of YAML tags, perl objects and data types. The inner API is not yet stable. Suggestions welcome.

You can check out all current parse and load results from the yaml-test-suite here:


You can alter the behaviour of YAML::PP by using the following schema classes:


One of the three YAML 1.2 official schemas


One of the three YAML 1.2 official schemas. Default


One of the three YAML 1.2 official schemas


Schema implementing the most common YAML 1.1 types


Serializing Perl objects and types


Serializing binary data


In progress. Keeping hash key order.


YAML 1.1 merge keys for mappings


Include other YAML files via !include tags

To make the parsing process faster, you can plugin the libyaml parser with YAML::PP::LibYAML.


The process of loading and dumping is split into the following steps:


    YAML Stream        Tokens        Event List        Data Structure
              --------->    --------->        --------->
                lex           parse           construct


    Data Structure       Event List        YAML Stream
                --------->        --------->
                represent           emit

You can dump basic perl types like hashes, arrays, scalars (strings, numbers). For dumping blessed objects and things like coderefs have a look at YAML::PP::Perl/YAML::PP::Schema::Perl.

For keeping your ordered Tie::IxHash hashes, try out YAML::PP::Schema::Tie::IxHash.


The Lexer is reading the YAML stream into tokens. This makes it possible to generate syntax highlighted YAML output.

Note that the API to retrieve the tokens will change.


The Parser retrieves the tokens from the Lexer. The main YAML content is then parsed with the Grammar.


The Constructor creates a data structure from the Parser events.


The Loader combines the constructor and parser.


The Dumper will delegate to the Representer


The Representer will create Emitter events from the given data structure.


The Emitter creates a YAML stream.


Still TODO:

Implicit collection keys
    [ a, b, c ]: value
Implicit mapping in flow syle sequences
    [ a, b, c: d ]
    # equals
    [ a, b, { c: d } ]
Plain mapping keys ending with colons
    key ends with two colons::: value
Supported Characters

If you have valid YAML that's not parsed, or the other way round, please create an issue.

Line and Column Numbers

You will see line and column numbers in the error message. The column numbers might still be wrong in some cases.

Error Messages

The error messages need to be improved.

Unicode Surrogate Pairs

Currently loaded as single characters without validating

Possibly more


The Constructor now supports all three YAML 1.2 Schemas, Failsafe, JSON and JSON. Additionally you can choose the schema for YAML 1.1 as YAML1_1.

Too see what strings are resolved as booleans, numbers, null etc. look at t/31.schema.t.

You can choose the Schema, however, the API for that is not yet fixed. Currently it looks like this:

    my $ypp = YAML::PP->new(schema => ['Core']); # default is 'JSON'

The Tags !!seq and !!map are still ignored for now.

It supports:

Handling of Anchors/Aliases

Like in modules like YAML, the Constructor will use references for mappings and sequences, but obviously not for scalars.

Boolean Handling

You can choose between 'perl' (1/'', currently default), 'JSON::PP' and 'boolean'.pm for handling boolean types. That allows you to dump the data structure with one of the JSON modules without losing information about booleans.


Numbers are created as real numbers instead of strings, so that they are dumped correctly by modules like JSON::PP or JSON::XS, for example.

See "NUMBERS" for an example.

Complex Keys

Mapping Keys in YAML can be more than just scalars. Of course, you can't load that into a native perl structure. The Constructor will stringify those keys with Data::Dumper instead of just returning something like HASH(0x55dc1b5d0178).


    use YAML::PP;
    use JSON::PP;
    my $ypp = YAML::PP->new;
    my $coder = JSON::PP->new->ascii->pretty->allow_nonref->canonical;
    my $yaml = <<'EOM';
                a: 1
                c: 2
            : 23
        : 42
    my $data = $yppl->load_string($yaml);
    say $coder->encode($data);
       "complex" : {
          "{'{a => 1,c => 2}' => 23}" : 42


Parse Tree

I would like to generate a complete parse tree, that allows you to manipulate the data structure and also dump it, including all whitespaces and comments. The spec says that this is throwaway content, but I read that many people wish to be able to keep the comments.

YAML::PP::Dumper, YAML::PP::Emitter

The Dumper should be able to dump strings correctly, adding quotes whenever a plain scalar would look like a special string, like true, or when it contains or starts with characters that are not allowed.

Most strings will be dumped as plain scalars without quotes. If they contain special characters or have a special meaning, they will be dumped with single quotes. If they contain control characters, including <"\n">, they will be dumped with double quotes.

It will recognize JSON::PP::Boolean and objects and dump them correctly.

Numbers which also have a PV flag will be recognized as numbers and not as strings:

    my $int = 23;
    say "int: $int"; # $int will now also have a PV flag

That means that if you accidentally use a string in numeric context, it will also be recognized as a number:

    my $string = "23";
    my $something = $string + 0;
    print $yp->dump_string($string);
    # will be emitted as an integer without quotes!

The layout is like libyaml output:

    - a
    - b
    - c
    - key1: 1
      key2: 2
      key3: 3
    - - a1
      - a2
    - - b1
      - b2


    my $ypp = YAML::PP->new;
    # load booleans via
    my $ypp = YAML::PP->new( boolean => 'boolean' );
    # load booleans via JSON::PP::true/false
    my $ypp = YAML::PP->new( boolean => 'JSON::PP' );
    # use YAML 1.2 Failsafe Schema
    my $ypp = YAML::PP->new( schema => ['Failsafe'] );
    # use YAML 1.2 JSON Schema
    my $ypp = YAML::PP->new( schema => ['JSON'] );
    # use YAML 1.2 Core Schema
    my $ypp = YAML::PP->new( schema => ['Core'] );
    # Die when detecting cyclic references
    my $ypp = YAML::PP->new( cyclic_refs => 'fatal' );
    # Other values:
    # warn   - Just warn about them and replace with undef
    # ignore - replace with undef
    # allow  - Default
    my $ypp = YAML::PP->new(
        boolean => 'JSON::PP',
        schema => ['JSON'],
        cyclic_refs => 'fatal',
        indent => 4, # use 4 spaces for dumping indentation
        header => 1, # default 1; print document header ---
        footer => 1, # default 0; print document footer ...
    my $doc = $ypp->load_string("foo: bar");
    my @docs = $ypp->load_string("foo: bar\n---\n- a");

Input should be Unicode characters (decoded).

    my $doc = $ypp->load_file("file.yaml");
    my @docs = $ypp->load_file("file.yaml");

Strings will be loaded as unicode characters (decoded).

    my $yaml = $ypp->dump_string($doc);
    my $yaml = $ypp->dump_string($doc1, $doc2);
    my $yaml = $ypp->dump_string(@docs);

Input strings should be Unicode characters. If not, they will be upgraded with utf8::upgrade.

Output will return Unicode characters (decoded).

    $ypp->dump_file("file.yaml", $doc);
    $ypp->dump_file("file.yaml", $doc1, $doc2);
    $ypp->dump_file("file.yaml", @docs);

Input data should be UTF-8 decoded. If not, it will be upgraded with utf8::upgrade.


This will dump to a predefined writer. By default it will just use the YAML::PP::Writer and output a string.

    my $writer = MyWriter->new(\my $output);
    my $yp = YAML::PP->new(
        writer => $writer,

Returns or sets the loader object, by default YAML::PP::Loader


Returns or sets the dumper object, by default YAML::PP::Dumper


Returns or sets the schema object


Creates and returns the default schema


The functions Load, LoadFile, Dump and DumpFile are provided as a drop-in replacement for other existing YAML processors. No function is exported by default.

    use YAML::PP qw/ Load /;
    my $doc = Load($yaml);
    my @docs = Load($yaml);

Works like load_string.

    use YAML::PP qw/ LoadFile /;
    my $doc = LoadFile($file);
    my @docs = LoadFile($file);
    my @docs = LoadFile($filehandle);

Works like load_file.

    use YAML::PP qw/ Dump /;
    my $yaml = Dump($doc);
    my $yaml = Dump(@docs);

Works like dump_string.

    use YAML::PP qw/ DumpFile /;
    DumpFile($file, $doc);
    DumpFile($file, @docs);
    DumpFile($filehandle, @docs);

Works like dump_file.


Compare the output of the following YAML Loaders and JSON::PP dump:

    use JSON::PP;
    use Devel::Peek;

    use YAML::XS ();
    use YAML ();
        $YAML::Numify = 1; # since version 1.23
    use YAML::Syck ();
        $YAML::Syck::ImplicitTyping = 1;
    use YAML::Tiny ();
    use YAML::PP;

    my $yaml = "foo: 23";

    my $d1 = YAML::XS::Load($yaml);
    my $d2 = YAML::Load($yaml);
    my $d3 = YAML::Syck::Load($yaml);
    my $d4 = YAML::Tiny->read_string($yaml)->[0];
    my $d5 = YAML::PP->new->load_string($yaml);

    Dump $d1->{foo};
    Dump $d2->{foo};
    Dump $d3->{foo};
    Dump $d4->{foo};
    Dump $d5->{foo};

    say encode_json($d1);
    say encode_json($d2);
    say encode_json($d3);
    say encode_json($d4);
    say encode_json($d5);

    SV = PVIV(0x55bbaff2bae0) at 0x55bbaff26518
      REFCNT = 1
      IV = 23
      PV = 0x55bbb06e67a0 "23"\0
      CUR = 2
      LEN = 10
    SV = PVMG(0x55bbb08959b0) at 0x55bbb08fc6e8
      REFCNT = 1
      FLAGS = (IOK,pIOK)
      IV = 23
      NV = 0
      PV = 0
    SV = IV(0x55bbaffcb3b0) at 0x55bbaffcb3c0
      REFCNT = 1
      FLAGS = (IOK,pIOK)
      IV = 23
    SV = PVMG(0x55bbaff2f1f0) at 0x55bbb08fc8c8
      REFCNT = 1
      FLAGS = (POK,pPOK,UTF8)
      IV = 0
      NV = 0
      PV = 0x55bbb0909d00 "23"\0 [UTF8 "23"]
      CUR = 2
      LEN = 10
    SV = PVMG(0x55bbaff2f6d0) at 0x55bbb08b2c10
      REFCNT = 1
      FLAGS = (IOK,pIOK)
      IV = 23
      NV = 0
      PV = 0



All the available parsers and loaders for Perl are behaving differently, and more important, aren't conforming to the spec. YAML::XS is doing pretty well, but libyaml only handles YAML 1.1 and diverges a bit from the spec. The pure perl loaders lack support for a number of features.

I was going over issues end of 216, integrating old patches from and creating some pull requests myself. I realized that it would be difficult to patch to parse YAML 1.1 or even 1.2, and it would also break existing usages relying on the current behaviour.

In 2016 Ingy döt Net initiated two really cool projects:


These projects are a big help for any developer. So I got the idea to write my own parser and started on New Year's Day 2017. Without the test suite and the editor I would have never started this.

I also started another YAML Test project which allows to get a quick overview of which frameworks support which YAML features:



It contains about 230 test cases and expected parsing events and more. There will be more tests coming. This test suite allows to write parsers without turning the examples from the Specification into tests yourself. Also the examples aren't completely covering all cases - the test suite aims to do that.

The suite contains .tml files, and in a seperate 'data' branch you will find the content in seperate files, if you can't or don't want to use TestML.

Thanks also to Felix Krause, who is writing a YAML parser in Nim. He turned all the spec examples into test cases.


This is a tool to play around with several YAML parsers and loaders in vim.

The project contains the code to build the frameworks (16 as of this writing) and put it into one big Docker image.

It also contains the yaml-editor itself, which will start a vim in the docker container. It uses a lot of funky vimscript that makes playing with it easy and useful. You can choose which frameworks you want to test and see the output in a grid of vim windows.

Especially when writing a parser it is extremely helpful to have all the test cases and be able to play around with your own examples to see how they are handled.


I was curious to see how the different frameworks handle the test cases, so, using the test suite and the docker image, I wrote some code that runs the tests, manipulates the output to compare it with the expected output, and created a matrix view.

You can find the latest build at

As of this writing, the test matrix only contains valid test cases. Invalid ones will be added.


Ingy döt Net

Ingy is one of the creators of YAML. In 2016 he started the YAML Test Suite and the YAML Editor. He also made useful suggestions on the class hierarchy of YAML::PP.

Felix "flyx" Krause

Felix answered countless questions about the YAML Specification.




The Perl Foundation sponsored this project (and the YAML Test Suite) with a grant of 2500 USD in 2017-2018.


Copyright 2018 by Tina Müller

This library is free software and may be distributed under the same terms as perl itself.