[Python] Shouldn't the argument name `include_key` of `polars.DataFrame.partition_by` be changed to `include_keys`? about polars HOT 4 CLOSED

Thank you for the explanation. I think that explanation makes sense.

from polars.

There is only one key. The key may be a tuple (i.e. a collection of columns), but a collection is still a singular item.

from polars.

There is only one key.

However, elsewhere it is written in the plurals as keys.

from polars.

This is a difference in usage. Imagine that the "key" to entering a secret club is a passcode, and each week that passcode changes. There is one key--the passcode--but the collection of passcodes form multiple key values, which one might refer to as "the keys."

In other words, there are two usages of the word "key" here: one is the key definition, the other is the key values. For the partition_by parameter, it makes sense to say "include the key associated with each element," because each element has only one key. It would also make sense to say "include the keys associated with the various values".

The second usage (the key values) would require us to determine how many unique keys there are. For example, the following dictionaries both have one key value

{
    "key1": "a",
}
{
    ("key1a", "key1b"): "a",
}

whereas the following has multiple keys:

{
    "key1": "a",
    "key2": "b",
}

Would you expect in the second case to use include_keys but include_key in the first case? It's more consistent and universally applicable to use the singular.

from polars.