Execute based on values of different data types in different columns of a data frame

I am trying to loop through all items in a column.

If it meets the condition in the column then check to see, If another condition is met in a different column.

The data types in the two columns are different, one is an integer & the other a string that I am checking for.

I am not sure where I have gone wrong & all information I've found online so far is about creating new columns based on data from a column.

Here is a simplified version of the code I am using.

import pandas as pd

servercount = 0
othercount = 0

df = pd.DataFrame({
    'operatingsystem': ['Windows 7', 'Windows Server 2012', 'Windows Server 2012 R', 
                'macOSX', 'Linux Server'],
    'login': [9, 12, 5, 11, 19],
})


for i in df['login']:
    if i > 10:
        if i in df['operatingsystem'].str.lower().str.contains("server"):
            servercount  = 1
        else:
            othercount  = 1

print("Servers:", servercount, "\nOthers:", othercount)

Currently this returns 0 on the servercount & 3 on the othercount meaning that it is not detecting the word server in the operatingsystem column.

CodePudding user response：

If you only want the final number (as in total servercount and othercount), then consider simply passing the logics to your dataframe. You are iterating over i only with is df['Login'], nowhere are you using operatingsystem values.

servercount = ((df['login'] > 10) & \
           (df['operatingsystem'].str.contains("server",
                                               case=False,
                                               regex=True))).sum()

Then, by difference you can calculate othercount:

othercount = len(df) - servercount

CodePudding user response：

Create mask wich chain both masks toghether by & bot bitwise AND and for count use Series.value_counts, last select True and False values:

m = (df['login'] > 10) & df['operatingsystem'].str.contains("server", case=False)
print (m.value_counts())
False    3
True     2
dtype: int64


out = m.value_counts()

print("Servers:", out[True], "\nOthers:", out[False])
Servers: 2 
Others: 3

EDIT: Because need 3 counts are created 3 mask separately, ~ is for invert mask (same like mask == False):

m1 = df['login'] > 10 
m2 =  df['operatingsystem'].str.contains("server", case=False)

print("Servers greater 10:", (m1 & m2).sum(), 
      "\nServers less 10:", (m1 & ~m2).sum(), 
      "\nOthers:", (~m1).sum())
Servers greater 10: 2 
Servers less 10: 1 
Others: 2