yaml

YAML 1.1 parser and emitter for MATLAB R2019b or newer. Based on SnakeYAML 1.30.

Examples

Load and dump

>> data.a = [1.23, 4.56];
>> data.b = {int32(2), {true, "hello", yaml.Null}};

>> s = yaml.dump(data)
    "a: [1.23, 4.56]
     b:
     - 2
     - [true, hello, null]
     "
   
>> result = yaml.load(s)
    a: {[1.2300]  [4.5600]}
    b: {[2]  {1×3 cell}}

Read and write files

>> yaml.dumpFile("test.yaml", data)
>> result = yaml.loadFile("test.yaml")
    a: {[1.2300]  [4.5600]}
    b: {[2]  {1×3 cell}}

Styles

>> s = yaml.dump(data, "auto")  % default
    "a: [1.23, 4.56]
     b:
     - 2
     - [true, hello, null]
     "
     
>> s = yaml.dump(data, "block")
    "a: 
     - 1.23
     - 4.56
     b:
     - 2
     - - true
       - hello
       - null
     "
     
>> s = yaml.dump(data, "flow")
    "{a: [1.23, 4.56], b: [2, [true, hello, 'null']], c: [2, [true, hola]]}
     "

YAML null

>> result = yaml.load("null")
    Null
    
>> yaml.isNull(result)
   1
   
>> s = yaml.dump(yaml.Null)
    "null
     "

Load YAML sequences as MATLAB standard arrays

By default, sequences are loaded as nested cell arrays to distinguish between YAML scalars and YAML one-element sequences and to supported mixed type sequences. If you use the ConvertToArray option, sequences are converted to 1D or 2D standard arrays if possible:

>> yaml.load("[[1, 2], [3, 4]]", "ConvertToArray", true)
     1     2
     3     4

Control dumping behaviour for MATLAB arrays

Since every MATLAB scalar is always an array and every array technically has at least 2 dimensions, there exists two ambiguities when dumping arrays:

MATLAB scalar → YAML scalar (default) or YAML one-element sequence
MATLAB vector → YAML sequence (default) or YAML sequence containing one YAML sequence

To avoid theses ambiguities and get consistent conversion behaviour, convert all your array data to nested vector cells before dumping them.

>> yaml.dump({1})
    "[1.0]
    "
>> yaml.dump({{1, 2}})
    "- [1.0, 2.0]
    "

Tinynja / yaml