CLN: Orderdict -> defaultdict(list)

[ShaharNaveh]

• ref CLN: OrderedDict -> Dict #30469
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[datapythonista]

The change is more complex than this. The code below can be simplified if this is a defaultdict.

Also, is OrderedDict still being used after this, or can it be removed from the imports?

[alimcmaster1]

Quite a few other uses of OrderedDict -> i'm aiming to eliminate most in this issue #30469

[ShaharNaveh]

Also, is OrderedDict still being used after this, or can it be removed from the imports?

Yes OrderedDIct is still being used.

I am really new to developing (I just greped for the words 'droped' '35' 'support' etc..), can you please explain what is the problem with OrderedDict ?

[datapythonista]

With a regular dict:

groupping_values = {}
for value in [1, 3, 1, 1, 2, 3]:
    if value not in groupping_values:
        groupping_values[value] = [value]
    else:
        groupping_values[value].append(value)

The if is needed, because when we create the key in the dictionary, we need to create the list with its value. The other times we can append to it.

With an example, if we have an empty dictionary {} and we want to add the Python programming language, we can do stuff['languages'] = ['python']. If now that we already have languages with a list, if we want to add javascript we use a different instruction stuff['languages'].append('javascript'). This is what the if in the example does.

Python defaultdict exists to make things simpler:

import collections
groupping_values = collections.defaultdict(list)
for value in [1, 3, 1, 1, 2, 3]:
    groupping_values[value].append(value)

A defaultdict is "smart", since in the creation we tell its elements will be list. So, when we call the append, it'll create an empty list automatically before. We don't need the if anymore, and the code is simpler.

This is what it's expected here.

[ShaharNaveh]

Thank you so much for the detailed explanation.

[datapythonista]

lgtm, thanks for cleaning this @MomIsBestFriend

[ShaharNaveh]

lgtm, thanks for cleaning this @MomIsBestFriend

Thanks for the detailed explanation, and the patience ;)

[jreback]

thanks @MomIsBestFriend