"试图在 Panda DataFrame 中洗牌"

Question

Hopefully, someone can help, I'm trying to randomize the output 15 times, and save into excel, however, python is only giving me 1 output instead of 15

import pandas as pd

# create a DataFrame
sales_to_do = {'Task': ['Call with the client', 'Preparing for the calls', 'Training staff',
                    'Daily tasks (Emails, questions, chasing)'],

           'Type of task': ['Call (external - new lead)', 'Preparing communication with 
leads', 'Training',
                            'Call (external - new lead)']}

df = pd.DataFrame(sales_to_do)
df_shuffled = df.sample(frac=1)

def randomize():
    df = pd.DataFrame(sales_to_do)
    df_shuffled = df.sample(frac=1)
    print(df_shuffled)


for i in range(15):
    randomize()

df_shuffled.to_excel(r'C:\Users\Alex\Desktop\Output1.xlsx', index=False, header=True)

Answer 1

You need to review the scoping rules. You have two independent variables named df_shuffled<\/code> , one each in randomize<\/code> and your main program. You never link the two. As a result, all that randomize<\/code> does is to shuffle the local DF and print the result -- the main program never references that ordering. At the end of your main, you simply dump the unchanged df_shuffled<\/code> from your first sampling.

def randomize():
    df = pd.DataFrame(sales_to_do)
    df_shuffled = df.sample(frac=1)
    return df_shuffled

for i in range(15):
    df_shuffled = randomize()
    # Adapt this output to append results per your needs
    df_shuffled.to_excel(r'C:\Users\Alex\Desktop\Output1.xlsx',
        index=False, header=True)

Answer 2

Something like this where you just return the shuffled df, and use pd.concat<\/code> on a list of these.

sales_to_do = pd.DataFrame({'id':[1,2], 'name':['bob','mike']})

def randomize(df):
    return df.sample(frac=1)
    
df_shuffled = pd.concat([randomize(sales_to_do) for x in range(15)])

df_shuffled.to_excel(r'C:\Users\Alex\Desktop\Output1.xlsx', index=False, header=True)

Answer 3

Currently the easiest and quick way to do it is using shuffle<\/code> from sklearn.utils<\/code>

    from sklearn.utils import shuffle
    df = shuffle(df, random_state=0)