Add axis=1 for concat #1009
Add axis=1 for concat #1009
Conversation
|
Nice! Thanks for working on this! |
|
haven't come up with a nice and robust solution for allowing |
Codecov Report
@@ Coverage Diff @@
## master #1009 +/- ##
=========================================
- Coverage 95.2% 95.1% -0.11%
=========================================
Files 34 34
Lines 6883 6779 -104
=========================================
- Hits 6553 6447 -106
- Misses 330 332 +2
Continue to review full report at Codecov.
|
| new_objs.append(obj) | ||
| objs = new_objs | ||
|
|
||
| if axis == 1: |
There was a problem hiding this comment.
For ignore_index, I think we can explicitly throw an exception for now when axis is 1.
databricks/koalas/namespace.py
Outdated
| if axis in [0, 'index']: | ||
| axis = 0 | ||
| elif axis in [1, 'columns']: | ||
| axis = 1 |
| sdfs = internal._sdf | ||
| continue | ||
| col = '{}'.format(index_col) | ||
| sdfs = sdfs.join(obj._internal._sdf, |
There was a problem hiding this comment.
@charlesdong1991, can we do this via each column assignment? It can leverage existing logics for compute.ops_on_diff_frames (see config.py)
There was a problem hiding this comment.
Very sorry to wait such long to get back to this PR @HyukjinKwon
I am not sure if i fully understand here, if via each column assignment, doesn't it mean the ops become more expensive? I made a slight change, feel free to review and will appreciate your feedback and suggestions! I got through tough time already, so will have time to make code changes much quicker.
|
Many thanks for your comments @HyukjinKwon I am very sorry that I have been very busy recently and don't have time for my existing PRs. I will try my best to update the PR following your comments this weekend! Thanks again for your understanding and patience! |
|
Hey, @charlesdong1991 no worries at all. Thanks for trying this change out. |
Softagram Impact Report for pull/1009 (head commit: f672805)⭐ Change Overview
📄 Full report
Impact Report explained. Give feedback on this report to [email protected] |
This PR proposes to add `axis` at `ks.concat(...)`. ```python import databricks.koalas as ks df1 = ks.DataFrame([['a', 1], ['b', 2]], columns=['letter', 'number']) df2 = ks.DataFrame([['bird', 'polly'], ['monkey', 'george']], columns=['animal', 'name']) ks.concat([df1, df2], axis=1) ``` ``` letter number animal name 0 a 1 bird polly 1 b 2 monkey george ``` Resolves #625, Closes #1009.
4 participants
#625