How to filter rows where columns meet consecutive conditions in Python?

Question

I’m trying to filter lines in which the columns comply with conditions consecutively. That is, if the row has columns with the conditions of after an L/I, the next column has a A/S, then return the value of 1 in the new column (if no, return 0)

Input:

       RFA RFB RFC RFD       
    0   S   S   S   S   
    1   A   I   A   A       
    2   A   A   L   A       
    4   S   S   L   A       

Output:

       RFA RFB RFC RFD  promo
    0   S   S   S   S     0
    1   A   I   A   A     1 
    2   A   A   L   A     1
    4   S   S   L   A     1

Script:

      def promo_behaviour(x):
          for i in range(0,95411):
             for j in data_rfa_r.columns:
                 if (x[j][i] == 'L' or x[j][i] == 'I') and (x[j][i+1] == 'A' or x[j][i+1] == 'S'):
                    return 1
                 else:
                    return 0
      data_rfa_r['promo'] = data_rfa_r.apply(promo_behaviour)

I wrote this function but without success (95411 are the number of remarks/lines).

I forgot to mention that in the context of the problem, the index column 0 is the latest! I mean, it should be read from right to left.

EDIT:

Output:

       RFA promo2 RFB promo1 RFC RFD    
    0   S    0    S     0     S   S   
    1   A    1    I     0     A   A     
    2   A    0    A     1     L   A   
    4   S    0    S     0     L   A   

   

Answer 1

You can use isin by creating a list of possible combinations.

vl = ['LA','IA','LS','IS']
dados['promo'] = (dados.shift(axis = 1) + dados).isin(vl).any(axis = 1).astype(int)

1. dados.shift 'move' the data frame
2. isin checks the occurrence within the list
3. any checks for occurrence of True on lines
4. astype(int) returns 0 or 1 instead of True or False

Answer 2

One Line Solution:

df['promo']=pd.Series([bool(re.search(r'(L|I)(?=[AS])',k)) for k in df.sum(axis=1)])

My idea was to transform the columns into a single column with the concatenation of the other columns. In this new column, I applied the logic test using regex. I used Positive lookbehind to check if there is an A or S after I saw an I or L. Whole code:

import pandas as pd
import numpy as np
import re

df=pd.read_csv("stack.txt",sep=",")

df['promo']=pd.Series([bool(re.search(r'(L|I)(?=[AS])',k)) for k in df.sum(axis=1)]).map({True:1, False:0})

print(df)

Returns:

  RFA RFB RFC RFD  promo
0   S   S   S   S      0
1   A   I   A   A      1
2   A   A   L   A      1
3   S   S   L   A      1