DOC: fix a failing doctest in DataFrame.to_dict #22827
Conversation
|
Hello @Moisan! Thanks for submitting the PR.
|
Codecov Report
@@ Coverage Diff @@
## master #22827 +/- ##
=======================================
Coverage 92.14% 92.14%
=======================================
Files 170 170
Lines 51017 51017
=======================================
Hits 47011 47011
Misses 4006 4006
Continue to review full report at Codecov.
|
pandas/core/frame.py
Outdated
| @@ -1122,7 +1123,7 @@ def to_dict(self, orient='dict', into=dict): | |||
| [{'col1': 1.0, 'col2': 0.5}, {'col1': 2.0, 'col2': 0.75}] | |||
|
|
|||
| >>> df.to_dict('index') | |||
| {'a': {'col1': 1.0, 'col2': 0.5}, 'b': {'col1': 2.0, 'col2': 0.75}} | |||
| {'a': {'col1': 1, 'col2': 0.5}, 'b': {'col1': 2, 'col2': 0.75}} | |||
There was a problem hiding this comment.
I'm ok merging this and just making the tests pass, but if you want to take care here of making the examples easier to understand, I think it'd be good as in the other strings to use real-world (e.g. animal legs...) examples.
From this, it takes a bit to see what's a, 1... Using something meaningful would be immediate.
There was a problem hiding this comment.
@datapythonista I agree with you in general :-), but here I think the 'col1' etc actually makes it easier to see all the different possible output formats (to clearly see where the column names go in the specific output type).
But 'a', 'b' is indeed less clear, but so I would rather call them 'row1', 'row2.
There was a problem hiding this comment.
That also makes sense. I'm personally more comfortable with something like {'cat': {'legs': 4, 'tail': 1}} than {'row1': {'col1': 0.75, 'col2': 1}}, but as far as it's not needed to spend a lot of unnecessary time until the example is understood, I'm happy with any option.
There was a problem hiding this comment.
Sorry to bother you again with the same @Moisan, but do you mind replacing the labels of the rows by row1 and row2, instead of a and b? I think that will make the example clearer
|
Sorry, from looking at the diff now, I don't find this an actual improvement. Eg in the output I really have to think hard "are legs/wings the columns or was it cat/falcon ?" to know what the structure is ( I find this is easier to see from a glance with something like: |
pandas/core/frame.py
Outdated
| @@ -1085,51 +1085,52 @@ def to_dict(self, orient='dict', into=dict): | |||
|
|
|||
| Returns | |||
| ------- | |||
| result : collections.Mapping like {column -> {index -> value}} | |||
| dict, list or collections.Mapping | |||
| like {column -> {index -> value}} | |||
There was a problem hiding this comment.
Sorry to ask you one more change. Do you mind changing the description of the Returns to something more accurate? That is only true with the default orient. Also, it'd be nice to start it with a capital letter and end it with a period. Thanks!
There was a problem hiding this comment.
No problem, better make it right during this pull request than having to come back later!
|
thanks @Moisan |
git diff upstream/master -u -- "*.py" | flake8 --diffBased on #22459. Fix the docstring for DataFrame.to_dict. I also updated
ci/doctests.sh.