read_csv return wrong dataframe when setting skiprows.

[strnam]

Code Sample, a copy-pastable example if possible

>>> import pandas as pd
>>> from StringIO import StringIO
>>> data = """id,text,num_lines
1,"line 11
line 12",2
2,"line 21
line 22",2
3,"line 31",1"""

>>> pd.read_csv(StringIO(data))
Out[2]: 
   id              text  num_lines
0   1  'line 11\nline 12'          2
1   2  'line 21\nline 22'          2
2   3           'line 31'          1

>>> pd.read_csv(StringIO(data), skiprows=[1])
Out[3]: 
         id              text  num_lines
0  'line 12"'                 2        NaN
1         2  'line 21\nline 22'        2.0
2         3           'line 31'        1.0
...

Expected Output

>>> pd.read_csv(StringIO(data), skiprows=[1])
Out[3]: 
   id              text  num_lines
0   2  'line 21\nline 22'          2
1   3           'line 31'          1
...

It should skip '1,"line 11\nline 12",2' instead skip '1,"line 11'

output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 2.7.10.final.0
python-bits: 64
OS: Linux
OS-release: 4.2.3-300.fc23.x86_64
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8

pandas: 0.18.0
nose: 1.3.7
pip: 8.1.1
setuptools: 18.0.1
Cython: None
numpy: 1.11.0
scipy: 0.14.1
statsmodels: 0.6.1
xarray: None
IPython: 3.2.1
sphinx: 1.2.3
patsy: 0.4.1
dateutil: 2.5.2
pytz: 2016.3
blosc: None
bottleneck: 0.6.0
tables: 3.2.2
numexpr: 2.4.6
matplotlib: 1.4.3
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: None
httplib2: None
apiclient: None
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: 2.8
boto: None

[jreback]

very tricky as you have embedded newlines within quoted fields. I guess the skip_lines is not accounting for the quoted fields (and ignoring them).

[jreback]

this is related to #10911

[jreback]

cc @mdmueller
cc @selasley
cc @evanpw

[strnam]

Thank you. That problem appear in my real project when other people give me a large csv file contain text column (text from news website). Because the file is big so I just want to read a part of file using skiprows parameter, but it don't work as I expect.

[selasley]

Using the python engine instead of the faster c engine works for the data given above

In [4]: pd.read_csv(StringIO(data), skiprows=[1], engine='python')
Out[4]: 
   id              text  num_lines
0   2  line 21\nline 22          2
1   3           line 31          1