Pandas breaks warnings module when printing DataFrame

[fisheye36]

>>> import warnings
>>> warnings.warn('asd')
<input>:1: UserWarning: asd
>>> warnings.warn('asd')
>>> warnings.warn('asd')
>>> import pandas as pd
>>> df = pd.DataFrame([1])
>>> warnings.warn('asd')
<input>:1: UserWarning: asd
>>> warnings.warn('asd')
>>> df
   0
0  1
>>> warnings.warn('asd')
<input>:1: UserWarning: asd
>>> warnings.warn('asd')
<input>:1: UserWarning: asd
>>> warnings.warn('asd')
<input>:1: UserWarning: asd
>>> warnings.warn('asd')
<input>:1: UserWarning: asd

Problem description

As you can see from the interactive Python listed above, when DataFrame is printed, suddenly all calls to warnings.warn() are being reported, instead of being suppressed when identical warning is issued.

The problem does not occur on Pandas version 0.25.3.

Output of pd.show_versions()

INSTALLED VERSIONS

commit : None
python : 3.8.1.final.0
python-bits : 64
OS : Windows
OS-release : 10
machine : AMD64
processor : Intel64 Family 6 Model 158 Stepping 10, GenuineIntel
byteorder : little
LC_ALL : None
LANG : None
LOCALE : English_United States.1250
pandas : 1.0.1
numpy : 1.18.1
pytz : 2019.3
dateutil : 2.8.1
pip : 20.0.2
setuptools : 45.1.0
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : 1.2.7
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : None
IPython : None
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : None
gcsfs : None
lxml.etree : None
matplotlib : 3.1.3
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pytables : None
pytest : None
pyxlsb : None
s3fs : None
scipy : None
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : 1.2.0
xlwt : None
xlsxwriter : 1.2.7
numba : None

[jorisvandenbossche]

Can you show the result of pd.show_versions() for both cases (for pandas 1.0 and pandas 0.25.3)? To check that there are no other differences between both environments

[jorisvandenbossche]

BTW, I can't reproduce this with pandas master on Python 3.7, and also not with pandas 1.0.0 on Python 3.8

[fisheye36]

@jorisvandenbossche, here is the same output for 0.25.3:

>>> import warnings
>>> import pandas as pd
>>> warnings.warn('asd')
<input>:1: UserWarning: asd
>>> warnings.warn('asd')
>>> warnings.warn('asd')
>>> df = pd.DataFrame([1])
>>> warnings.warn('asd')
>>> warnings.warn('asd')
>>> warnings.warn('asd')
>>> df
   0
0  1
>>> warnings.warn('asd')
>>> warnings.warn('asd')
>>> warnings.warn('asd')

Output of pd.show_versions() for 0.25.3:

INSTALLED VERSIONS

commit : None
python : 3.8.1.final.0
python-bits : 64
OS : Windows
OS-release : 10
machine : AMD64
processor : Intel64 Family 6 Model 158 Stepping 10, GenuineIntel
byteorder : little
LC_ALL : None
LANG : None
LOCALE : English_United States.1250

pandas : 0.25.3
numpy : 1.18.1
pytz : 2019.3
dateutil : 2.8.1
pip : 20.0.2
setuptools : 45.1.0
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : None
IPython : None
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : None
gcsfs : None
lxml.etree : None
matplotlib : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pytables : None
s3fs : None
scipy : None
sqlalchemy : None
tables : None
xarray : None
xlrd : None
xlwt : None
xlsxwriter : None

[fisheye36]

A little update.

First of all, all the tests were being done using PyCharm's Python Console. When I wrote a script file and then run it, the behavior is a bit different. Here is the script:

import sys
import warnings
from time import sleep

import pandas as pd


def do_warn():
    sleep(1)
    warnings.warn('asd')
    sys.stdout.flush()


print('Printing warning 3 times', flush=True)
do_warn()
do_warn()
do_warn()

print('Creating DataFrame and then printing warning 3 times', flush=True)
df = pd.DataFrame([1])
do_warn()
do_warn()
do_warn()

print('Printing DataFrame and then printing warning 3 times', flush=True)
print(df)
do_warn()
do_warn()
do_warn()

print('Printing DataFrame AGAIN and then printing warning 3 times', flush=True)
print(df)
do_warn()
do_warn()
do_warn()

print('Printing DataFrame AGAIN and then printing warning 3 times', flush=True)
print(df)
do_warn()
do_warn()
do_warn()

I use sleep() to align stdout and stderr, otherwise stderr was being printed after all stdout messages (which is strange, I thought flush=True should fix that but that is not the point of this issue). Here is the output:

C:\Users\fisheye\Documents\Projects\pandastest\venv\Scripts\python.exe C:/Users/fisheye/Documents/Projects/pandastest/src/test.py
Printing warning 3 times
C:/Users/fisheye/Documents/Projects/pandastest/src/test.py:10: UserWarning: asd
  warnings.warn('asd')
Creating DataFrame and then printing warning 3 times
Printing DataFrame and then printing warning 3 times
   0
0  1
C:/Users/fisheye/Documents/Projects/pandastest/src/test.py:10: UserWarning: asd
  warnings.warn('asd')
Printing DataFrame AGAIN and then printing warning 3 times
   0
0  1
C:/Users/fisheye/Documents/Projects/pandastest/src/test.py:10: UserWarning: asd
  warnings.warn('asd')
Printing DataFrame AGAIN and then printing warning 3 times
   0
0  1
C:/Users/fisheye/Documents/Projects/pandastest/src/test.py:10: UserWarning: asd
  warnings.warn('asd')

Process finished with exit code 0

It seems like printing DataFrame somehow resets warnings altogether. In case of PyCharm's Python Console, after just one print it keeps resetting warnings every time, maybe due some internal printing of that DataFrame somewhere in the background (a list of variables maybe?).

[mroeschke]

Seems like this may be a PyCharm issue potentially since it's not reproducible with other set ups so closing. Can reopen if it appears again