ENH: series rank has a percentage rank option #5978
Conversation
|
can you add a vbench for both cases (the original and new one?) |
| @@ -940,6 +940,7 @@ New features | |||
| - Access to historical Google Finance data in pandas.io.data (:issue:`3814`) | |||
| - DataFrame plotting methods can sample column colors from a Matplotlib | |||
| colormap via the ``colormap`` keyword. (:issue:`3860`) | |||
| - ``Series.rank()`` has a percentage rank option (:issue: `5971`) | |||
There was a problem hiding this comment.
I think you put it accidentally in pandas 0.12 section
There was a problem hiding this comment.
Thanks, that is fixed.
|
any thoughts on 'percentage' vs. say 'per'? @jorisvandenbossche @y-p ? |
|
I prefer 'pct' to 'per' @jorisvandenbossche @y-p |
|
@MichaelWS so change to that.. |
|
done |
|
also, I have never worked with vbench before, so I am not sure if I am doing what is intended |
|
add a benchmark in vb_suite/groupby.py.....just copy-paste what you want to do
|
|
I am getting some odd errors. I will have to do some more research later today on it. sqlalchemy.exc.IntegrityError: (IntegrityError) column checksum is not unique u'INSERT INTO benchmarks (checksum, name, description) VALUES (?, ?, ?)' ('ea1993ef61c3cc4e871d2cce3c5d983c', 'eval_frame_chained_cmp_python', None) |
| @@ -3894,6 +3894,11 @@ def test_rank(self): | |||
| iranks = iseries.rank() | |||
| exp = iseries.astype(float).rank() | |||
| assert_series_equal(iranks, exp) | |||
| iseries = Series(np.arange(5)) + 1.0 | |||
There was a problem hiding this comment.
can you add a couple of more tests, maybe all nan series and for groupby, group that has 1 element
There was a problem hiding this comment.
added more tests with partial nan's and duplicate values. nan's will always be nan's so not sure if we would ever catch a bug if all nan's.
There was a problem hiding this comment.
but that is the check; make sure you propogate nans; the edge cases are always important to test (and usually the hardest to get right)
There was a problem hiding this comment.
easy enough. I added that as well.
|
this looks fine |
|
easy enough. That is done. |
| @@ -79,9 +79,12 @@ Improvements to existing features | |||
| - ``plot(legend='reverse')`` will now reverse the order of legend labels for most plot kinds. | |||
| (:issue:`6014`) | |||
| - Allow multi-index slicers (:issue:`6134`, :issue:`4036`, :issue:`3057`, :issue:`2598`, :issue:`5641`) | |||
| <<<<<<< HEAD | |||
There was a problem hiding this comment.
your rebase introduced these....need to edit
|
can you reset the release.rst to master HEAD then add your change again. Their were some formatting updates which you are reverting....this should only be your 1 line change |
ENH: series rank has a percentage rank option
|
thanks @MichaelWS ! |
|
Thanks @jreback |
|
@MichaelWS I think we need to have this issue is here: #6717 |
closes #5971
This pull allows people to compute percentage ranks in cython for a series. I found myself computing this all the time and it will make this calculation much faster.