Map keys is broken about pydash

[zbynekcz12]

Yes, your analysis is correct. The map_keys() function in Pydash iterates over the object's keys and values in the order (value, key), which is the opposite of the order that most iterators use. This is because the map_keys() function needs to know the key in order to generate the new key.

However, this order can cause problems if the iterator function expects the values to be strings. For example, the str.upper() function expects a string argument, but it will raise a TypeError if it is passed an integer.

To fix this, Pydash could change the order of the arguments in the iteriteratee() function to (key, value). This would make the map_keys() function more consistent with other iterators and would also prevent the error that you are seeing.

In the meantime, there are a few workarounds that you can use. One option is to explicitly pass the key argument to the iterator function. For example:

from pydash import py_

def upper_key(key):
    return key.upper()

assert py_(dict(a=1, b=2, c=3)).map_keys(upper_key).value() == dict(A=1, B=2, C=3)

Another option is to use a lambda function that takes the key and value arguments in the desired order. For example:

from pydash import py_

assert py_(dict(a=1, b=2, c=3)).map_keys(lambda key, value: key.upper()).value() == dict(A=1, B=2, C=3)

Finally, you could also use a list comprehension to iterate over the object's keys and values and generate the new keys. For example:

from pydash import py_

assert py_({key.upper(): value for key, value in dict(a=1, b=2, c=3).items()}).value() == dict(A=1, B=2, C=3)

I hope this helps!

[Drvanon]

Thank you for your quick response! I would like to point out that both your snippets fail:

assert py_(dict(a=1, b=2, c=3)).map_keys(lambda key: key.upper()).value() == dict(A=1, B=2, C=3)

fails because the key argument is an integer. Your second snippet also does not work, since key and value need to be swapped:

from pydash import py_

assert py_(dict(a=1, b=2, c=3)).map_keys(lambda _value, key: key.upper()).value() == dict(A=1, B=2, C=3)

I agree that your third option does indeed work, however I really enjoy the way that I can use postfix notation on chains. It, so to say, breaks the chain.

[dgilland]

iteriteratee calls like most iteratees with the values first and the keys second, but that is undesirable behavior for this function.

The behavior of map_keys has always been to pass (value, key, object) to the callback or the first N arguments if the callback accepts fewer than 3 arguments:

pydash/src/pydash/objects.py

Lines 1183 to 1187 in 64d4614

	def map_keys(obj, iteratee=None):
	"""
	The opposite of :func:`map_values`, this method creates an object with the same values as object
	and keys generated by running each own enumerable string keyed property of object through
	iteratee. The iteratee is invoked with three arguments: ``(value, key, object)``.

This mirrors the behavior of Lodash's mapKeys: https://lodash.com/docs/4.17.15#mapKeys