Add columns in pandas dataframe dynamically

Question

I have following code to load dataframe

import pandas as pd

ufo = pd.read_csv('csv_path')
print(ufo.loc[[0,1,2] , :])

which gives following output, see the structure of the csv

          City Colors Reported Shape Reported State             Time
0       Ithaca             NaN       TRIANGLE    NY   6/1/1930 22:00
1  Willingboro             NaN          OTHER    NJ  6/30/1930 20:00
2      Holyoke             NaN           OVAL    CO  2/15/1931 14:00

Now, I want to add an extra column based on existing column. I have a list which consist of indices of participating columns. It can be 0,1 or 0,2,3 or 1,2,3 anything.

I need to create it dynamically. I could come up with following

df1['combined'] = df1['City']+','+df1['State']

Putting index doesn't seem to work. I want to join those columns. using ','.join()

akuiper · Accepted Answer · 2016-11-07 14:21:56Z

3

Assuming the data types of all the columns you want to join are str, you can use [] with integer to pick up the columns and use apply to join them:

df[[0,2,3]].apply(','.join, axis=1)

#0      Ithaca,TRIANGLE,NY
#1    Willingboro,OTHER,NJ
#2         Holyoke,OVAL,CO
#dtype: object

answered Nov 7, 2016 at 14:21

akuiper

216k33 gold badges362 silver badges379 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Ami Tavory · Accepted Answer · 2016-11-07 14:24:56Z

3

If the list of indices is l, you can use pd.Series.cat:

df[df.columns[l[0]]].astype(str).str.cat([df[df.columns[i]].astype(str) for i in l[1: ]], sep=',')

Example

In [18]: df = pd.DataFrame({'a': [1, 2], 'b': [2, 'b'], 'c': [3, 'd']})

In [19]: df[df.columns[l[0]]].astype(str).str.cat([df[df.columns[i]].astype(str) for i in l[1: ]], sep=',')
Out[19]: 
0    1,2
1    2,b
Name: a, dtype: object

answered Nov 7, 2016 at 14:24

Ami Tavory

76.7k13 gold badges152 silver badges196 bronze badges

2 Comments

mdurant Over a year ago

I believe this is the same: df[[l[0]]].str.cat(df[l[1:]], sep=',')

Ami Tavory Over a year ago

@mdurant Thanks. I suspect you might be right in principle, but your specific code (and some of its variants) simply don't run by me. The first problem I get is 'DataFrame' object has no attribute 'str' (but fixing this gives me other errors).

piRSquared · Accepted Answer · 2016-11-07 14:41:43Z

3

def dyna_join(df, positions):
    return pd.concat([df, df.iloc[:, positions].apply(','.join, 1).rename('new_col')], axis=1)


dyna_join(df, [0, -2])

answered Nov 7, 2016 at 14:41

piRSquared

296k68 gold badges509 silver badges654 bronze badges

1 Comment

jezrael Over a year ago

and maybe add astype(str) if int or float column.

Collectives™ on Stack Overflow

Add columns in pandas dataframe dynamically

3 Answers 3

Comments

2 Comments

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

2 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related