Flexible and powerful data analysis / manipulation library for Python, providing labeled data structures similar to R data.frame objects, statistical functions, and much more
Today I had to use DataFrame.lookup on a dataframe of +- 1.5 million rows and 30 columns and found it wasn't very performant, then I looked at the code and saw it's a for loop internally.
for i, (r, c) in enumerate(zip(row_labels, col_labels)):
result[i] = self._get_value(r, c)I created my own method with DataFrame.melt and loc, which performed better:
def lookup(data, col_label):
result = data.melt(id_vars=col_label)
result = values.loc[values[col_label] == values['variable'], 'value']
return resultWould it be an idea to open a ENH ticket for this? Thought I check it here first.
Deprecation warning should precede extended summary. I tried to move it around in the docstring (starting at the top of the docstring), but still get the error when I try to validate here. I checked some other example where deprecation warning is added, and they look similar to mine. Can someone point me in the right direction here? Thanks
Deprecation warning should precede extended summary at #35224, the docstring error does not make lot of sense to me, so not sure how to solve it
1 reply
Here's the current str_get implementation
def f(x):
if isinstance(x, dict):
return x.get(i)
elif len(x) > i >= -len(x):
return x[i]
return np.nanCan we change it to something like this:
def f(x):
try:
return x[i]
except (KeyError, IndexError):
return np.nanpd.Series.str.get documentation states parameters as:
Parameters: i: int
Position of element to extract.i can be any hashable object right? In the case of dictionaries i need not necessarily be anint. Would it be good to change it to i: hashable object?
17 replies
import pandas as pd; pd.Timestamp("2019-10-27 02:00:00+02:00", tz="Europe/Warsaw").ceil(freq="1H") erroring out with pytz.exceptions.AmbiguousTimeError: Cannot infer dst time from 2019-10-27 02:00:00, try using the 'ambiguous' argument? I would expect the .ceil to be a no-op here. Pandas 1.0.5
For a PR to be merged is it a requirement that all of the Checks pass? Or can a PR be merged with some of the Checks failing?
I ask because some tests are failing on py37_np_dev; looks like they are failing for every PR.
Is it still worth chasing PR for approval/merge while the tests are in this failing state or should we wait til these tests are fixed?
@jreback @WillAyd @rhshadrach I'm sure any one of you could answer.
1 reply
1 reply
9 replies
cyberjunkie Spamming ads
3 replies
