String methods get() does not insert NaNs on negative index out of range

[louise-davies]

Code Sample, a copy-pastable example if possible

import pandas as pd
df1 = pd.DataFrame([["1_2_3_4_5"], ["6_7_8_9_10"], ["11_12"]], columns=["test"])

# access using positive index
df1["test"].str.split("_").str[2]
# output:
# 0        3
# 1        8
# 2      NaN

# access using negative index
df1["test"].str.split("_").str[-3]
# output:
# IndexError: list index out of range

Problem description

Since accessing an array using a negative index is supported if all items in the Series are the correct length for it, and if you access using a positive index that is out of range for any item in the series instead outputs a NaN for that item, I would think that the behaviour should be unified and that access using a negative index also outputs NaNs if an item in the Series is shorter than expected.

Expected Output

df1["test"].str.split("_").str[-3]
# output:
# 0        3
# 1        8
# 2      NaN

Output of pd.show_versions()

[paste the output of pd.show_versions() here below this line]
INSTALLED VERSIONS

commit: None
python: 3.4.3.final.0
python-bits: 64
OS: Linux
OS-release: 3.13.0-132-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8
LOCALE: en_US.UTF-8

pandas: 0.20.3
pytest: None
pip: 9.0.1
setuptools: 36.2.7
Cython: None
numpy: 1.13.1
scipy: 0.19.1
xarray: None
IPython: 6.1.0
sphinx: None
patsy: None
dateutil: 2.6.1
pytz: 2017.2
blosc: None
bottleneck: None
tables: None
numexpr: None
feather: None
matplotlib: 2.0.2
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: None
html5lib: 0.999999999
sqlalchemy: 1.1.14
pymysql: None
psycopg2: None
jinja2: 2.9.6
s3fs: None
pandas_gbq: None
pandas_datareader: None

[sinhrks]

Thx for the report. Yeah we should fix it to check negative boundary. PR is welcome.

pandas/pandas/core/strings.py

Line 1263 in f6d4d70

f = lambda x: x[i] if len(x) > i else np.nan

[bobhaffner]

I can submit a PR for this