Skip to content


Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP


merge with empty DataFrame IndexError #1628

idoko opened this Issue · 2 comments

4 participants

In [1]: from pandas import *

In [2]: a = DataFrame(columns=['column1'], data=[1])

In [3]: b = DataFrame(columns=['column1'])

In [4]: a.merge(b, on=['column1'], how='left')

IndexError                                Traceback (most recent call last)
IndexError: index out of range for array
@wesm wesm closed this in 0da0066
@yarikoptic yarikoptic referenced this issue from a commit in neurodebian/pandas
@yarikoptic yarikoptic Merge tag 'v0.8.1' into debian
Version 0.8.1

* tag 'v0.8.1': (126 commits)
  RLS: Version 0.8.1
  DOC: tweak
  DOC: set_index/reset_index examples
  DOC: doc fixes and what's new in 0.8.1, vectorized string methods
  ENH: better string element access/slicing notation close #1656
  DOC: minor additions to release notes for 0.8.1
  BUG: handle Yahoo! finance returning duplicate dates for prev bus day, doc fixes
  BUG: fix windows/32-bit builds
  BUG: get #1620 fix working on python 3
  ENH: handling of UTF-8 strings in DataFrame columns, close #1620
  TST: span unit test #1635
  TST: skip another @network test if no internet connection
  ENH/BUG: handle tz-aware datetime.datetime in to_datetime, add utc=True option to allow conversion to utc, close #1581
  ENH: hack to not compress single group keys, accelerate single-key and Categorical groupby operations
  BUG: fix merge bug with left joins on length-0 DataFrame, close #1628
  BUG: Series.interpolate bug with method='values' and datetime64[ns], close #1646
  BUG: properly handle None values in dict input to concat, close #1649
  BUG: len-0 Series min/max/describe #1650
  Fix describe() failure for None and empty Series.
  BUG: string date aliases now work with tz-aware time series close #1647

Still getting this error on pandas 0.7


Upgrade not possible?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Something went wrong with that request. Please try again.