Python Pandas Calculation of a Column with conditions-CodePudding

I'm trying to make a calculation on 2 columns using python pandas. Use case is like this:

I have values like 100,101,102... in column "hesapKodu1". I have split this column in 3 columns for the first 3 characters. "hesapkodu1_1" is the first character of "hesapKodu1" so it is "1". "hesapKodu1_2" is the first 2 characters of "hesapkodu1", so it is like "10,11"...

What I'm trying to do is this: When hesapKodu1 is 123 or 125 or 130 I would like to make calculation for columns BORC and ALACAK: it will be like BORC - ALACAK.

But for the other hesapKodu1 values it will be ALACAK - BORC.

And at the end all of the results should be summed as a return.

Right now my code is like this. And this code can only do BORC - ALACAK when hesapKodu1 starts with 1. I cannot find a way to iterate through upper conditions.

source_df['hesapKodu1_1']=source_df['hesapKodu1'].str[:1]
source_df['hesapKodu1_2']=source_df['hesapKodu1'].str[:2]
source_df['hesapKodu1_3']=source_df['hesapKodu1'].str[:3]
hk1 = round(source_df.loc[source_df['hesapKodu1_1'] == '1', 'BORÇ'].sum() - source_df.loc[source_df['hesapKodu1_1'] == '1', 'ALACAK'].sum(),2)
h

CodePudding user response：

You can make use of np.where() which would be faster than apply().

import numpy as np

source_df["new_column"] = np.where(
    source_df["hesapKodu1"].isin(["123", "125", "130"]),
    source_df["BORC"] - source_df["ALACAK"],
    source_df["ALACAK"] - source_df["BORC"],
)

CodePudding user response：

You can use apply for that -

def my_func(record):
    if source_df['hesapKodu1'] in ['123', '125', '130']
       record['new_column'] = record['BORC']  - record['ALACAK']
    else:
       record['new_column'] = record['ALACAK']  - record['BORC']
    return record

target_df = source_df.apply(my_func, axis=1)