YAML and tuples as dict key

Question

Let

d = {3: u'blah', u'b': u'foo', (12,2,3) : u'bonjour'}

be a dictionary. If I use

g = yaml.dump(d)
h = yaml.load(g)
print g, h

it's very good, but I get lots of !!python/unicode everywhere, so that's not good for readability.

On the other hand, if I use :

g = yaml.safe_dump(d)
h = yaml.load(g)
print g, h

there are no-more the !!python/unicode but also not the !!python/tuple anymore, so the load doesn't work because the tuple isn't understood correctly.

How to get the right balance with YAML ?

• no !!python/unicode everywhere,
• having tuples as dict key working ! (so probably !!python/typle has to be kept)

Answer 1

If you search for python yaml unicode, the very first result is this bug report, which has exactly the solution you're looking for. In short, register a custom representer for unicode strings, like this:

yaml.add_representer(unicode, 
                     lambda dumper, value: dumper.represent_scalar(u'tag:yaml.org,2002:str', value))

With this in place, your example renders like this:

>>> d = {3: u'blah', u'b': u'foo', (12,2,3) : u'bonjour'}
>>> print yaml.dump(d)
3: blah
? !!python/tuple [12, 2, 3]
: bonjour
b: foo