Pandas: ENH: Support mangle_dupe_cols=False in pd.read_csv()
12935 added full support for duplicate column names (in header or in names) by mangling them. While this has been considered _acceptable_ by users, ideally, we would like to not have to mangle them.
gfyoung
All 7 comments
@jreback : Given what you said in #17060, is this something we should still pursue ?
gfyoung
on 27 Jul 2017
Depending on how difficult this is, I would personally still have it as our goal to have mangle_dupe_cols=False implemented some time.
jorisvandenbossche
on 31 Jul 2017
What is the ETA on this issue?
caniko
on 19 Sep 2018
when / if a community pull request happens
jreback
on 19 Sep 2018
@caniko2 : This is quite a tricky one given that duplicate column names have unusual behavior in pandas. You are more than welcome to submit a PR to implement it if you like.
gfyoung
on 19 Sep 2018
Could anyone help me to check whether current pandas 0.24.2 support "mangle_dupe_cols=False"?
I find docs at http://pandas.pydata.org/pandas-docs/stable/user_guide/io.html, showing : Passing in False will cause data to be overwritten if there are duplicate names in the columns.
Thanks so much!
jackzhenguo
on 22 May 2019
Still no support, as behavior of data handling has proven to be quite non-trivial when there are duplicate column names. You are welcome to give it a shot though!
gfyoung
on 22 May 2019
Related issues
matthiasroder
路
3Comments
tade0726
路
3Comments
Ashutosh-Srivastav
路
3Comments
mfmain
路
3Comments
nathanielatom
路
3Comments
Most helpful comment
Depending on how difficult this is, I would personally still have it as our goal to have
mangle_dupe_cols=Falseimplemented some time.