BUG: DataFrame.apply(raw=True) applies function twice on first row/column

[alonme]

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

Code Sample, a copy-pastable example

In [1]: import pandas as pd
   ...:
   ...: df = pd.DataFrame({'a': [1,2,3]})
   ...:
   ...: def mul2(x):
   ...:     print('Hello')
   ...:     return x*2
   ...:
   ...: print('Apply:')
   ...: df.apply(mul2)
   ...:
   ...: print('\nApply Raw:')
   ...: df.apply(mul2, raw=True)

Apply:
Hello

Apply Raw:
Hello
Hello

Problem description

Because of an implementation detail, the applied function is applied twice on the first row/column,
This results in side effects happening twice.

Other than the apply_raw code path, this issues should be fixed by #34183.

Expected Output

Apply:
Hello

Apply Raw:
Hello

Output of pd.show_versions()

INSTALLED VERSIONS

commit : 07edc87
python : 3.7.5.final.0
python-bits : 64
OS : Darwin
OS-release : 19.5.0
Version : Darwin Kernel Version 19.5.0: Thu Apr 30 18:25:59 PDT 2020; root:xnu-6153.121.1~7/RELEASE_X86_64
machine : x86_64
processor : i386
byteorder : little
LC_ALL : en_US.UTF-8
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.1.0.dev0+1706.g07edc871f.dirty
numpy : 1.18.3
pytz : 2020.1
dateutil : 2.8.1
pip : 19.2.3
setuptools : 41.2.0
Cython : 0.29.17
pytest : 5.4.1
hypothesis : 5.11.0
sphinx : 3.0.3
blosc : 1.9.1
feather : None
xlsxwriter : 1.2.8
lxml.etree : 4.5.0
html5lib : 1.0.1
pymysql : None
psycopg2 : None
jinja2 : 2.11.2
IPython : 7.14.0
pandas_datareader: None
bs4 : 4.9.0
bottleneck : 1.3.2
fastparquet : 0.3.3
gcsfs : None
matplotlib : 3.2.1
numexpr : 2.7.1
odfpy : None
openpyxl : 3.0.3
pandas_gbq : None
pyarrow : 0.17.0
pytables : None
pyxlsb : None
s3fs : 0.4.2
scipy : 1.4.1
sqlalchemy : 1.3.16
tables : 3.6.1
tabulate : 0.8.7
xarray : 0.15.1
xlrd : 1.2.0
xlwt : 1.3.0
numba : 0.49.0

[arw2019]

The reason why you see the side effect twice is because in pandas/core/apply.py's apply_raw function

pandas/pandas/core/apply.py

Line 221 in 48c00ea

def apply_raw(self):

mul2 gets called twice, first inside the try statement with pandas._libs's compute_reduction, and then after an error is thrown, in the except block.

I haven't understood the error in compute_reduction here entirely. I tracked it down as far as this function call:

pandas/pandas/_libs/reduction.pyx

Line 629 in 48c00ea

reducer = Reducer(arr, f, axis=axis, dummy=dummy, labels=labels)

However, my impression from the comments apply_raw is that it's thrown intentionally, and apply_raw is working as expected.

I'm happy to work on a pull request/documentation update if we think there are changes or improvements to be made!

[alonme]

Hey @arw2019, I am familiar,
I solved this issue except for raw=True in #34183
Hope it will be merged soon, and then not much would be left to be done in apply_raw

[alonme]

@arw2019 My PR was merged #34183
You can look at how i solved the non-raw path, raw can work in a very similar fashion.

[arw2019]

@alonme Cool! I'll work on a pull request and loop you in!

[alonme]

Per #34913, and comments on that issue, i think we can close this issue.

[simonjayhawkins]

on master

>>> df = pd.DataFrame({"a": [1, 2, 3]})
>>>
>>>
>>> def mul2(x):
...     print("Hello")
...     return x * 2
...
>>>
>>> print("Apply:")
Apply:
>>> df.apply(mul2)
Hello
   a
0  2
1  4
2  6
>>>
>>> print("\nApply Raw:")

Apply Raw:
>>> df.apply(mul2, raw=True)
Hello
   a
0  2
1  4
2  6
>>> pd.__version__
'1.2.0.dev0+314.g44e933a765'
>>>