DataFrame.to_csv with empty dataframe probleme quoting

[parisni]

Code Sample, a copy-pastable example if possible

>>>import pandas as pd
>>> pd.DataFrame([""]).to_csv("/tmp/test.csv", header= False, index= False, quoting=3)
[...]Error: single empty field record must be quoted

Problem description

I cannot get an empty csv. There is allways quote for the first row

Expected Output

I should get a csv with an empty row.

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 3.6.0.final.0 python-bits: 64 OS: Linux OS-release: 4.8.0-45-generic machine: x86_64 processor: x86_64 byteorder: little LC_ALL: None LANG: fr_FR.UTF-8 LOCALE: fr_FR.UTF-8

pandas: 0.19.2
nose: 1.3.7
pip: 9.0.1
setuptools: 27.2.0
Cython: 0.25.2
numpy: 1.12.0
scipy: 0.18.1
statsmodels: 0.6.1
xarray: None
IPython: 5.1.0
sphinx: 1.5.2
patsy: 0.4.1
dateutil: 2.6.0
pytz: 2016.10
blosc: None
bottleneck: 1.2.0
tables: 3.3.0
numexpr: 2.6.1
matplotlib: 2.0.0
openpyxl: 2.4.1
xlrd: 1.0.0
xlwt: 1.2.0
xlsxwriter: 0.9.6
lxml: 3.7.2
bs4: 4.5.3
html5lib: None
httplib2: 0.10.3
apiclient: 1.6.2
sqlalchemy: 1.1.5
pymysql: None
psycopg2: None
jinja2: 2.9.4
boto: 2.45.0
pandas_datareader: None

[jreback]

Is this meant to be something?

In [3]: pd.DataFrame([""])
Out[3]: 
  0
0  

In [2]: pd.DataFrame([]).to_csv(None, header= False, index= False, quoting=3)
Out[2]: ''

In [6]: pd.DataFrame(['foo']).to_csv(None, header= False, index= False, quoting=2)
Out[6]: '"foo"\n'

[jreback]

cc @gfyoung

[jreback]

this actually looks ok to me.

[parisni]

How should I get an empy csv with 2 rows ?
from [None, None]

cf http://stackoverflow.com/questions/20665141/writing-a-pandas-dataframe-into-a-csv-file-with-some-empty-rows

[parisni]

pd.DataFrame([None, None]).to_csv(None, header= False, index= False)
'""\n\n'

[jreback]

In [4]: pd.DataFrame(index=[0,1]).to_csv(None, header= False, index= False)
Out[4]: '\n\n'

None is an actual value.

[parisni]

Thank, this way solves my problem.
However, I cannot anderstand what the logic behind.
pd.DataFrame([None, None]).to_csv(None, header= False, index= False)
should return
'""\n""\n'
then

[jreback]

so the missing values are getting filled by the na_rep. But they are still quoted, which seems a bit odd. (as when I put an actual value they are NOT quoted).

In [6]:  pd.DataFrame([None, None])
Out[6]: 
      0
0  None
1  None

In [7]: pd.DataFrame([None, None]).to_csv(None, header= False, index= False, na_rep='foo')
Out[7]: 'foo\nfoo\n'

In [8]: pd.DataFrame([None, None]).to_csv(None, header= False, index= False, na_rep='')
Out[8]: '""\n\n'

[jreback]

I think this is a bug, so marked. @parisni welcome to have a PR to fix! (code is actually not bad for writing csvs....)

[mroeschke]

This looks fixed on master. Could use a test

In [3]: In [8]: pd.DataFrame([None, None]).to_csv(None, header= False, index= False, na_rep='')
Out[3]: '""\n""\n'