Skip to content
This repository was archived by the owner on May 4, 2019. It is now read-only.

Conversation

@simonster
Copy link
Member

As @garborg noticed, since #102, sorting of NAs in PDAs is inconsistent with DataArrays. This fixes PDAs to sort NAs to the end.

I put a nalast parameter on groupsort_indexer to control whether NAs are sorted first or last, as currently the grouping functions and tests in DataFrames seem to expect NAs in the first position. This should be revisited, but I need to go through that code to understand how it works (or maybe @powerdistribution can say what's best?).

I also moved the sorting tests from DataFrames here, and I now run the same sorting tests I run on DataArrays on PooledDataArrays. Sorting PooledDataArrays in reverse is broken, but that was a pre-existing bug.

@coveralls
Copy link

Coverage Status

Coverage decreased (-0.19%) when pulling 334229c on sjk/pdasort into 1359a53 on master.

This can be done more efficiently, but working is better than not
working.
@coveralls
Copy link

Coverage Status

Coverage decreased (-0.02%) when pulling 8479c13 on sjk/pdasort into 1359a53 on master.

@coveralls
Copy link

Coverage Status

Coverage decreased (-0.0%) when pulling 011f419 on sjk/pdasort into 1359a53 on master.

@coveralls
Copy link

Coverage Status

Coverage decreased (-0.0%) when pulling b5dee5f on sjk/pdasort into 1359a53 on master.

simonster added a commit that referenced this pull request Jul 10, 2014
Sort NAs to last position for PooledDataArrays as well
@simonster simonster merged commit 696ae34 into master Jul 10, 2014
@simonster simonster deleted the sjk/pdasort branch July 10, 2014 23:04
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants