A line in a self define function not working in pandas-CodePudding

I try and defined a function to process a df (like adding columns and convert all cols head to lower case) before doing the analysis. All other line works fine except the line that I tried to rearrange the columns orders.

the function looks like this

def cleanDf(df):    
    df.columns = df.columns.str.replace(' ','_')
    df.columns = df.columns.str.lower()
    df['date1'] = pd.to_datetime(df['date'].astype(str)   ' '   df['time'].astype(str))
    df['weekday'] =  df['date1'].dt.day_name()
    
    business_hour_mask = (df['date1'].dt.hour >=9) & (df['date1'].dt.hour <=18)
    df['business_hour'] = np.where(business_hour_mask, "Yes","No")
    df['week_number'] = df.date1.dt.week
    
    df = df.reindex(['date1','week_number','weekday','business_hour','changed_by','customer','field_name','new_value','old_value','new_value.1','old_value.1','date','time','company_code','sales_organization','distribution_channel','division'], axis=1) 
    #problem line, i've tried both with and without "df = " in front of this line
    
    return df

my current workaround is to insert that line after i call the function then it works

cleanDf(df)

df = df.reindex(['date1','week_number','weekday','business_hour','changed_by','customer','field_name','new_value','old_value','new_value.1','old_value.1','date','time','company_code','sales_organization','distribution_channel','division'], axis=1)

df.head()

Appreciate if you can advise why the line does not inside the function, but ok when executed separately.

thank you very much

CodePudding user response：

It's because you're reassigning the df variable inside the function, where it's just a parameter. Since you're returing df though, it's simple. Just write df = cleanDf(df) instead of just cleanDf(df):

df = cleanDf(df)
df.head()

Per @mozway's comment, you should also define your cleanDf function like so:

def cleanDf(df):
    df = df.copy()
    # ... do your stuff ...
    return df