How to fill (based on the index of a dataframe) an empty column-CodePudding

I'm trying to add the column 'Information' to my dataframe (df3) and filling it with string values ('True' if the index is 0 and 'False', otherwise). The problem is pandas put 'False' in every single row, even in the ones having an index 0 (see the output below).

Input :

import pandas as pd
import numpy as np

df1 = pd.DataFrame({'Column1': ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J'],
                    'Column2': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
                    'Column3': ['I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X'],
                    'Column4': ['K', 'L', 'M', 'N', 'O', 'P', 'Q', 'R', 'S', 'T'],
                    'Column5': [11, 12, 13, 14, 15, 16, 17, 18, 19, 20],
                    'Column6': ['XI', 'XII', 'XIII', 'XIV', 'XV', 'XVI', 'XVII', 'XVIII', 'XIX', 'XX'],
                    'Column7': ['U', 'V', 'W', 'X', 'Y', 'Z', '', '', '', ''],
                    'Column8': [21, 22, 23, 24, 25, 26, pd.NA, pd.NA, pd.NA, pd.NA],
                    'Column9': ['XXI', 'XXII', 'XXIII', 'XXIV', 'XXV', 'XXVI', '', '', '', '']})

column_names = ['Letters', 'Numbers', 'RomanNumerals']
df3 = pd.DataFrame(columns = column_names)

i=0
while i<len(df1.columns):
    df2 = df1.iloc[:, i:i 3]
    df2.columns = column_names
    df3 = pd.concat([df3, df2])
    i =3

df3.dropna(inplace=True)

for index, row in df3.iterrows():
    df3['Information'] = np.where(index == 0, True,  False)

display(df3)

Output :

	Letters	Number	RomanNumeral	Information
0	A	1	I	FALSE
1	B	2	II	FALSE
2	C	3	III	FALSE
3	D	4	IV	FALSE
4	E	5	V	FALSE
5	F	6	VI	FALSE
6	G	7	VII	FALSE
7	H	8	VIII	FALSE
8	I	9	IX	FALSE
9	J	10	X	FALSE
0	K	11	XI	FALSE
1	L	12	XII	FALSE
2	M	13	XIII	FALSE
3	N	14	XIV	FALSE
4	O	15	XV	FALSE
5	P	16	XVI	FALSE
6	Q	17	XVII	FALSE
7	R	18	XVIII	FALSE
8	S	19	XIX	FALSE
9	T	20	XX	FALSE
0	U	21	XXI	FALSE
1	V	22	XXII	FALSE
2	W	23	XXIII	FALSE
3	X	24	XXIV	FALSE
4	Y	25	XXV	FALSE
5	Z	26	XXVI	FALSE

Is there an explanation to this scenario ?

CodePudding user response：

Change the for loop with this snippet


df3['Information']= df3.index.map(lambda x: x==0)

What happen in the for loop is you actually make a new column based on a scalar. Not that you typed


df3['Information'] = np.where(index == 0, True,  False)

Instead of


row['Information'] = np.where(index == 0, True,  False)

But even the code above won't work because you assign to nothing

Edit:

Another way to do this (for further explanation you can check pandas dataframe apply)


def get_information(index):
    if index==0:
        return True
    else:
        return False

df3['Information']= df3.index.map(get_information)