assigning value to a column by applying if statement with multiple conditions-CodePudding

How can I fix this code where I'm trying to assign the values in a column according to the various conditions. Writing the below code gives me an error saying:

The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all()

list = [df1,df2,df3,df4]             # multiple dataframes
grp_list = ["con", "eco", "dip", "pol"]      # multiple categories in a column 

for i in list: 
    if i['pgp'].isin(group_list) and (i.egp == i.pgp):
        i['value'] == 1
    elif ~i['pgp'].isin(group_list):
        i['value'] == 2
    else:
        i['value'] == 0

Expected Output df1:

pgp     egp     value
con     con      1     # return 1 if pgp value is in the element list & pgp = egp
eco     eco      1     
dip     health   0     # else 0
pol     health   0
god     con      2
ent     eco      2     # return 2 if pgp value is not in the element list

CodePudding user response：

Use np.select~

def apply_conditions(df, group_list):
    in_group_list = df['pgp'].isin(group_list)
    conditions = (
        in_group_list & df['egp'].eq(df['pgp']),
        ~in_group_list
    )
    choices = (
        1,
        2
    )
    df['value'] = np.select(conditions, choices, default=0)
    return df

dfs = [df1,df2,df3,df4]
grp_list = ['con', 'eco', 'dip', 'pol']

dfs = [apply_conditions(df, grp_list) for df in dfs]

CodePudding user response：

You can use np.select here:

import numpy as np

df1 = pd.DataFrame({
    'pgp': ['con', 'eco', 'dip', 'pol', 'god', 'ent'],
    'egp': ['con', 'eco', 'health', 'health', 'con', 'eco']
})

grp_list = ["con", "eco", "dip", "pol"]

m1 = df1['pgp'].isin(grp_list)
m2 = df1['pgp'] == df1['egp']

conditions=[m1&m2, ~m1]
choices = [1, 2]

df1['value'] = np.select(conditions, choices, default=0)
print(df1)

Output:

   pgp     egp  value
0  con     con      1
1  eco     eco      1
2  dip  health      0
3  pol  health      0
4  god     con      2
5  ent     eco      2