python yaml load formatted string

Question

I'm trying to load yaml that contains python formatted strings, e.g. test: {formatted_string}. This would allow me to format the string using dictionary["test"].format(formatted_string="hello yaml"), but when I load the yaml, it's automatically converted to {'test': {'formatted_string': None}} instead of {'test': '{formatted_string}'}.

There are dozens of .yaml files that are already formatted in this way.

I don't see this in the pyyaml docs or anywhere on SO.

Code in full for clarity:

import yaml


data = """
test: {formatted_string}
"""
d1 = yaml.load(data)
print(d1)
# {'test': {'formatted_string': None}}

d2 = {"test": "{formatted_string}"}
print(d2)
# {'test': '{formatted_string}'}

d2["test"] = d2["test"].format(formatted_string="hello yaml")
print(d2)
# {'test': 'hello yaml'}

Thanks!

Answer 1

The { character in YAML (as in JSON) introduces a dictionary. That is this:

a_dictionary:
  key1: value1
  key2: value2

Is completely equivalent to:

a_dictionary: {key1: value1, key2: value2}

So when you write...

test: {formatted_string}

...the YAML parser things you are introducing a dictionary, and that it has a single key (formatted_string) and no value. If you want to use a { as part of a YAML value you need to quote it:

test: "{formatted_string}"

Compare:

>>> yaml.safe_load('test: {formatted_string}')
{'test': {'formatted_string': None}}
>>> yaml.safe_load('test: "{formatted_string}"')
{'test': '{formatted_string}'}

In general, if you always quote your YAML strings your life will be easier :).