[FIXED] Python Pandas dataframe : Fill dataframe with 0/1

Issue

My goal is to define a data frame matrix (df2) whose rows correspond to communes and the columns to the cantons of Switzerland. I need to fill in the matrix with 0/1 values where entry (i,j) is a 1 if the commune in row i is in the canton in column j and a 0 otherwise.

[df1] (parameters) : https://i.stack.imgur.com/0u7w8.png

[df2] (dataframe to fill) : https://i.stack.imgur.com/AR0mM.png

What i try to do is to loop like this :

for row in df2:
    for column in df2.columns:
        if column == df1.loc[row]:
            df2[column] = 1
        else:
            df2[column] = 0

But it doesn't run. The error is about "df1.loc[row]"

• KeyError: 'ZH'

Would appreciate any educational help. Thanks in advance :)

Solution

You can use the first dataframe to generate a crosstab or a pivot table, then use that to reindex like df2:

out = pd.crosstab(df1.index, df1['Canton']).reindex_like(df2)

or:

out = (df1
 .reset_index()
 .pivot_table(index='Commune', columns='Canton', aggfunc=lambda x: 1, fill_value=0)
 .reindex_like(df2)
)

example output:

Canton    BS  GE  VD  ZH
Commune                 
Basel      1   0   0   0
Genève     0   1   0   0
Lausanne   0   0   1   0
Zürich     0   0   0   1

Answered By - mozway