REGR: to_csv(na_rep) truncates values

[mdruiter]

Code Sample

>>> import pandas as pd
>>> pd.__version__
'1.0.0rc0'
>>> pd.Series(range(8, 12)).to_csv(na_rep='-')
',0\r\n0,8\r\n1,9\r\n2,1\r\n3,1\r\n'

Problem description

Despite (or due to?) #29975 and #30146, using na_rep still makes to_csv() truncate values. In this case integers.

Expected Output

',0\r\n0,8\r\n1,9\r\n2,10\r\n3,11\r\n'

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 142 Stepping 12, GenuineIntel byteorder : little LC_ALL : None LANG : None LOCALE : Dutch_Netherlands.1252

pandas : 1.0.0rc0
numpy : 1.18.1
pytz : 2019.3
dateutil : 2.8.1
pip : 20.0.2
setuptools : 41.2.0
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 2.10.3
IPython : 7.11.1
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : None
gcsfs : None
lxml.etree : None
matplotlib : 3.1.2
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pytables : None
pytest : None
s3fs : None
scipy : 1.4.1
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlwt : None
xlsxwriter : None
numba : None

[jorisvandenbossche]

Thanks for the report!

Yes, the itemsize = writers.word_len(na_rep) in the PR you linked seems like a wrong fix. Further investigations always welcome! I tagged it with the 1.0.1 milestone, as this is also a regression compared to 0.25 (#29975 was a new issue)

[fwitte]

Hi,
this is happening similarly to me, but in my case it affects pure boolean columns:

import pandas as pd
pd.__version__
'1.0.0'
pd.Series([True, False, 3.1113, 150, -10]).to_csv('outfile.csv', na_rep='nan')

correctly yields

,0
0,True
1,False
2,3.1113
3,150
4,-10

while

pd.Series([True, False, False, True]).to_csv('outfile.csv', na_rep='nan')

results in this:

,0
0,Tru
1,Fal
2,Fal
3,Tru

Output of pd.show_versions()

commit : None
python : 3.6.7.final.0
python-bits : 64
OS : Linux
OS-release : 4.15.0-76-generic
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : None
LANG : de_DE.UTF-8
LOCALE : de_DE.UTF-8

pandas : 1.0.0
numpy : 1.16.2
pytz : 2019.1
dateutil : 2.8.0
pip : 20.0.2
setuptools : 41.0.0
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : 2.8.4 (dt dec pq3 ext lo64)
jinja2 : 2.10.3
IPython : None
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : None
gcsfs : None
lxml.etree : None
matplotlib : 3.0.3
numexpr : 2.7.1
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pytables : None
pytest : None
pyxlsb : None
s3fs : None
scipy : 1.4.1
sqlalchemy : 1.3.11
tables : 3.6.1
tabulate : None
xarray : 0.14.1
xlrd : None
xlwt : None
xlsxwriter : None
numba : None

Have a nice weekend!

[TomAugspurger]

Taking a look at this now.

[TomAugspurger]

I think https://github.com/pandas-dev/pandas/pull/30146/files#diff-8f51ce8697bdff680b3f71b196c582ee is the bug. That will truncate the values to the length of na_rep, regardless of the length of the string repr values.