Replace use of dataframe with pandas . apply

Question

I have a function that iterates a spreadsheet, which makes some controls. I looked over and saw about . apply, but I couldn’t implode. I would like to know if you could in that case, replace the use I use and for performance purposes, change to . apply.

In this case, I have a dictionary that he consults for my treated conditions. And the function that iterates making a loop to make this happen. The function depends on the position I pass in the dictionary and that comes from the spreadsheet to know if it displays a photo field or not.

    for i, row in df.iterrows():
        if row['cargo'] in cargos_to_display_photo:
            df.loc[i, 'display_foto'] = True
        else:
            df.loc[i, 'display_foto'] = False
    return df 

the function that "prepares" the data frame is:

def prepare_dataframe(df_google):
    df_google = set_display_photo(df_google)
    return df_google

the dictionary:

cargos_to_display_photo = [
    'Trainee | Assessoria de Investimentos',
    'Sócio Fundador',
    'Superintendente',
    'Assessor de Investimentos',
    'Sócia | Assessor de Investimentos',
    'Líder Corporate',
    'Trainee | Corporate',
]

the column I consult:

columns_dict = {
    'SEXO' : 'sexo',
    'CÓDIGO' : 'codigo',
    'Nome Completo' : 'nome',
    'Cidade' : 'unidade',
    'Cargo' : 'cargo',
}

The columns that do not query or that were not cited in the excerpt I discard by the code itself, giving a drop in them with the necessary negotiations so that do not need to iterate.

the ways I researched, I couldn’t implode any. I couldn’t get apply to populate the spreadsheet. It would be possible to do this with these functions?

Answer 1

I looked over and saw about . apply, but I couldn’t implement it. I would like to know if you could in that case, replace the use I use and for performance purposes, change to . apply.

Importing the file:

import pandas as pd

df = pd.read_excel('./Cadastro Colaboradores BT.xlsx')

List of posts:

cargos_to_display_photo = [ 'Trainee | Assessoria de Investimentos', 
                           'Sócio Fundador', 'Superintendente', 
                           'Assessor de Investimentos', 
                           'Sócia | Assessor de Investimentos', 
                           'Líder Corporate', 'Trainee | Corporate', ]

Code with the apply:

df['display_foto'] = df['Cargo'].apply(lambda x: x in cargos_to_display_photo)

I create a column called display_photo, apply it to the Job column of the dataframe and make an anonymous function that checks the contents of the column row by row.

• Run time with for loop:

  69.4 ms 1.52 ms per loop (Mean Std. dev. of 7 runs, 10 loops each)

• Run time with apply:

  305 µs 30.8 µs per loop (Mean Std. dev. of 7 runs, 1000 loops each)

Answer 2

One way to do this would be by using the function .isin() pandas. It returns a boolean list when it finds some value within the array cargos_to_display_photo

df['display_foto'] = df['Cargo'].isin(cargos_to_display_photo)
df.head()
    SEXO    CÓDIGO  Nome completo   Cidade      Cargo       display_foto
0   m       NaN     Samuel...       Ribeir...   Sócio ...   False
1   m       NaN     Leonar...       Uberlâ...   Assess...   True
2   m       NaN     Andre ...       Franca      Sócio ...   False
3   f       NaN     Rejane...       Uberlâ...   Assess...   False
4   m       NaN     Wevers...       Rio de...   Traine...   True

A good practice to get performance(speed) in pandas runs is to avoid using repetition loops as well as function apply (which is also a type of iteration). Follow the implementation time of my proposal:

220 µs 5.93 µs per loop (Mean Std. dev. of 7 runs, 1000 loops each)