Concatenate columns containing lists by index

Asked Apr 27 '20 at 13:17

Active Apr 27 '20 at 13:17

Viewed 18 times

I have a logically simple problem but am going round in circles at the moment and could really appreciate some help.

I have ONE DataFrame which has as an index "datetime" objects. Furthermore, I have one column called "data" that contains one list of floats:

                 data
    timestamp
    27.04.2020   [1.1,  2.2]   
    27.04.2020   [3.3,  4.4]
    27.04.2020   [5.5,  6.6]
    26.04.2020   [7.7,  8.8]
    26.04.2020   [9.9,  10.10]

I now want everything to be merged by timestamp. So the 1st to 3rd row should be merged into a single list and so on:

                 data
    timestamp
    27.04.2020   [1.1, 2.2, 3.3, 4.4, 5.5, 6.6]   
    26.04.2020   [7.7, 8.8, 9.9, 10.10]

This is a minimal sample, I have 3 or more equal timestamps that need to be merged.

I have looked through many different topics that have been opened but couldn't find one that deals with one single DataFrame and lists.

Every help is appcriated,

Rising

asked Apr 27 '20 at 13:17

rising02

`df.groupby('timestamp').sum()`? – Quang Hoang Apr 27 '20 at 13:19
Perhaps `df.groupby(level=0).sum()` @quang – yatu Apr 27 '20 at 13:20
@yatu if index has a name, either way would work. – Quang Hoang Apr 27 '20 at 13:21
U right, it should work too being a named index @quang – yatu Apr 27 '20 at 13:22
It works with sum... but how and why? How does pandas know that the two lists should not be summed item by item but rather merged? – rising02 Apr 27 '20 at 13:33
It's not pandas. In python you can sum lists to *concatenate* them: try `[1,2,3]+[4,5,6]` – yatu Apr 27 '20 at 14:46
ah okay, thanks you very much :) – rising02 Apr 28 '20 at 08:49

Concatenate columns containing lists by index

0 Answers0