Pandas Dataframe Filtering with optional conditions

user_12383795 0
Viewed 93

I have a question about filltering pandas dataframe by optionals as below:

Request: in case 'brand'=='All" or 'lo' == 'All', will exclude from filtering. Issue: I try to process as below but doesn't work. Please help me to fix code.

def filter_for_company(df_source, season, brand, lo):
    mask = (
               (df_source['Production Priority'] == 'Primary')
            &  (df_source['Season'] == season)
            &  (if brand =='All':
                    pass
                else:
                    df_source['Brand'] == brand
                )
            &  (if lo ='All':
                    pass
                else:    
                    df_source['Liaison Office Code'] == lo
                )             
    )            
    company_df = df_source.loc[mask,:]
    return company_df
pandas dataframe python
2 Answers

The conditional slicing for pandas dataframes using loc or iloc are designed for filtering based on values actually present in the dataframe. What you want to achieve should be handled separately by the code which calls your filtering function.
However, if you do not have a control over that, you can modify your function as below:

def filter_for_company(df_source, season, brand, lo):
    mask = (
        (df_source["Production Priority"] == "Primary") 
        & (df_source["season"] == season)
    )

    if brand != "All":
        df_source = df_source[df_source["Brand"] == brand]

    if lo != "All":
        df_source = df_source[df_source["Liaison Office Code"] == lo]

    company_df = df_source.loc[mask, :]
    return company_df
user_53822130

Your if-else statement within the mask definition is not giving you the required boolean condition you want. You can try modifying the codes to properly set the boolean conditions, as follows:

def filter_for_company(df_source, season, brand, lo):
    mask = (
               (df_source['Production Priority'] == 'Primary')
            &  (df_source['Season'] == season)
            &  (brand !='All')
            &  (lo !='All')
    )            

    if brand !='All':
        df_source['Brand'] == brand

    if lo !='All':
        df_source['Liaison Office Code'] == lo
               
    company_df = df_source.loc[mask,:]
    return company_df

Note that this may not yet be the best optimized codes to achieve your purpose. However, without the overall picture of the dataframe and sample data, it is hard to further optimize it.

user_150706970
Related