The common element between a row and a column in a python dataframe

Question

I have two dataframes. The FIRST one, shown below, has three columns.

Col_1   Col_2   Col_3
aaa     dfd     ccc
sdf     jjj     sge
rty     fgh     rtg
hji     dfg     hyt
lkj     bgh     dcf

In each row, there is one element that is the same as one of the elements in the SECOND dataframe shown below (the elements in the second dataframe do not have to have any a specific order, of course).

list
ccc
sge
fgh
dfg
dcf

My goal is to iterate through each row in the FIRST dataframe and find that common element with the SECOND dataframe. This is followed by bringing that element ahead to the beginning of the row. The expected result is as follows:

Expected result

Col_1   Col_2   Col_3
ccc     aaa     dfd
sge     sdf     jjj
fgh     rty     rtg
dfg     hji     hyt
dcf     lkj     bgh

Any help will be appreciated !!

Answer 1

Using the .apply method of the pandas DataFrame you can do it in one line. This will be faster than manually iterating over the rows.

It only uses pandas and works at a row level by first checking if any of the rows elements are in ls, sorts the returned binary indicator (True to front of row) and then re-indexes the row to be sorted in this order. It then broadcasts the results back onto the original row.

import pandas as pd

df = pd.DataFrame({'col1':['aaa','sdf','rty','hji','lkj'],
                   'col2':['dfd','jjj','fgh','dfg','bgh'],
                   'col3':['ccc','sge','rtg','hyt','dcf']})     

ls = pd.Series(['ccc','sge','fgh','dfg','dcf'])

df = df.apply(lambda x: x[(~x.isin(ls)).argsort()], 
              axis=1, 
              result_type='broadcast')

Returns:

    col1  col2   col3
0   ccc    aaa   dfd
1   sge    sdf   jjj
2   fgh    rty   rtg
3   dfg    hji   hyt
4   dcf    lkj   bgh

Answer 2

Why not try using apply, isin and tolist:

print(df.apply(lambda x: x[x.isin(ls)].tolist() + x[~x.isin(ls)].tolist(), axis=1))

Output:

  col1 col2 col3
0  ccc  aaa  dfd
1  sge  sdf  jjj
2  fgh  rty  rtg
3  dfg  hji  hyt
4  dcf  lkj  bgh

I simple load each row and get the one that is in ls and make it the first value, by adding the rest to the end with itself being the first, using isin and tolist and +.