Owo #1
Owo #1
Conversation
| relabeling = func is None and is_multi_agg_with_relabel(**kwargs) | ||
| if relabeling: | ||
| # OWO CHANGES |
There was a problem hiding this comment.
We should make this into a new normalize_keyword_aggregation
There was a problem hiding this comment.
While we're at it, check if second field of tuple is a lambda, since native aggregation methods wouldn't work obviously
| @@ -393,7 +406,16 @@ def _agg(arg, func): | |||
| """ | |||
| result = {} | |||
| for fname, agg_how in arg.items(): | |||
| result[fname] = func(fname, agg_how) | |||
| # OWO CHANGES | |||
There was a problem hiding this comment.
Perhaps encapsulate the changes in a new agg function to keep some consistency (agg_multi_d1? or something)
pandas/core/groupby/generic.py
Outdated
| relabeling = func is None and is_multi_agg_with_relabel(**kwargs) | ||
| if relabeling: | ||
| # OWO CHANGES | ||
| import json |
There was a problem hiding this comment.
We can explore other tools like np.array().toBytes() (if more efficient) and also use the python list_like check thing
There was a problem hiding this comment.
Depends on if JSON is too slow and if they can't cythonize this code
| items = json.loads(fname) | ||
| _obj = {} | ||
| for item in items: | ||
| _obj[item] = self._gotitem(item, ndim=1, subset=None) |
There was a problem hiding this comment.
This is kinda bad... but hopefully with a formal PR, we can get feedback on how to do this better (maybe pipeline method would work better if we are to keep internal consistency)
pandas/core/base.py
Outdated
| @@ -393,7 +406,16 @@ def _agg(arg, func): | |||
| """ | |||
| result = {} | |||
| for fname, agg_how in arg.items(): | |||
| result[fname] = func(fname, agg_how) | |||
| # OWO CHANGES | |||
| import json | |||
There was a problem hiding this comment.
might have some issues with json performance that we need to discuss
| @@ -0,0 +1,81 @@ | |||
| { | |||
There was a problem hiding this comment.
might want to delete this or concert it into a testcase
|
Missing documentation for this feature, so I'll add this. |
black pandasgit diff upstream/master -u -- "*.py" | flake8 --diff