Convert column_names from a dataframe to a dictionary

Question

I have a dataframe like below:

Column1    Column2    Column3    Column4

1             12          5           6

I want to create a dictionary with the column_name as keys i.e. Column1, Column2, Column3, Column4 and values as 1,2,3,4 (incremental values)

Post that, i want to transform my dataframe to below:

1             2          3             4

1             12          5           6

i.e. Column1 replaced by 1, Column2 replaced by 2 and so on......

Can someone help me out?

https://stackoverflow.com/questions/26716616/convert-a-pandas-dataframe-to-a-dictionary — Kallz, Feb 06 '18 at 05:39
If you just need last character as column names, `df.columns = df.columns.str[-1]` — Zero, Feb 06 '18 at 05:44
I don't need the last characters, I just gave the column names as examples..i need to save the column names as a dictionary and replace the dataframe's column names with the dictionary values — Satyajit Pattnaik, Feb 06 '18 at 05:57

Tai · Answer 1 · 2018-02-06T05:51:09.987

1

IIUC, you can try this

d = {k: v for k, v in zip(df.columns, range(1, len(df.columns)+1))}    
Out[372]:
{'Column1': 1, 'Column2': 2, 'Column3': 3, 'Column4': 4}

After getting d, you can do

df.columns = df.columns.to_series().map(d).values

    1   2   3   4
0   1   12  5   6

Here we make use of map and turn df.columns into a pd.Series object.

edited Feb 06 '18 at 05:51

answered Feb 06 '18 at 05:43

Tai

7,684
3
29
49

1

Didn't the op ask for some output without 'Column'? – Stephen Rauch Feb 06 '18 at 05:44
@StephenRauch I am trying to follow the title...but good point. – Tai Feb 06 '18 at 05:47

score 1 · Answer 2 · answered Feb 06 '18 at 05:52

No need for a dictionary, you can assign range objects to df.columns directly.

df.columns = range(1, len(df.columns) + 1)

df
   1   2  3  4
0  1  12  5  6

You can also use any of the .str methods on df.columns -

df.columns = df.columns.str.lstrip('Column')

Or,

df.columns = df.columns.str.replace('^Column', '')

df
   1   2  3  4
0  1  12  5  6

score 0 · Answer 3 · answered Feb 06 '18 at 06:03

I think this codes will help you out:

import pandas as pd
import numpy as np
data = [{'column1': 1, 'column2': 12,'column3': 15, 'column4': 6}]
df = pd.DataFrame(data)
df.columns = range(1, len(df.columns) + 1)
print (df)

#the output goes like this :

df
   1   2  3  4
0  1  12  5  6

romulomadu · Answer 4 · 2018-02-06T09:26:56.033

You can do this:

import pandas as pd   
data = [{'column1': 1, 'column2': 12,'column3': 15, 'column4': 6}]
df = pd.DataFrame(data)
#dict of columns {'column1': 1, 'column2': 2, ...}
column_dict = dict(zip(df.columns, list(range(len(df.columns))+1)))
# rename dataframe columns using rename method
new_df = df.rename(columns=column_dict)

Or this:

# create a df without specify columns names 
# will generate columns as numbers from 0 to len(df.columns) - 1
new_df = pd.DataFrame(df.values)
# add 1 to starts by 1
new_df.columns = new_df.columns + 1

Convert column_names from a dataframe to a dictionary

4 Answers4