Home > Software design >  pandas: apply random.shuffle() to list column
pandas: apply random.shuffle() to list column

Time:04-29

I have a dataframe as follows,

import pandas as pd
df= pd.DataFrame({'text':['The weather is nice','the house is amazing','the flowers are blooming']})

I would like to shuffle the words in each row using random.shuffle(),(e.g the new first row will be 'nice is weather the' ),so I have done the following,

df.new_text = df.text.str.split()

and tried to map or apply shuffle() function but it returns None.

print(df.new_text.map(lambda x: random.shuffle(x)))

or

print(df.new_text.apply(lambda x: random.shuffle(x)))

I am not sure what I am doing wrong here. and then finally I would like to join the shuffled words in the list to get a string per row,

df.new_text = df.new_text.apply( lambda x:' '.join(x))

CodePudding user response:

This does the job.

shuffled_sentences = {"text":[]}

for sentence in df.values.ravel():
  np.random.shuffle(sentence)
  shuffled_sentences["text"].append(sentence)

shuffled_df = pd.DataFrame(shuffled_sentences)

The thing with np.random.shuffle is that it doesn't return any output. So you need to store the list you want to shuffle in a vraible first. Then if you apply np.random.shuffle on it, the original variable itself would be shuffled.

CodePudding user response:

You can use np.random.permutation from numpy library

df= pd.DataFrame({'text':['The weather is nice','the house is amazing','the 
flowers are blooming']})

df['new_text']= df['text'].apply(lambda x:x.split())
df['new_text']= df['new_text'].map(lambda x:  np.random.permutation(x))
df['new_text']= df['new_text'].apply( lambda x:' '.join(x))


display(df.new_text)
0         nice weather is The
1        the is house amazing
2    flowers are the blooming

CodePudding user response:

The problem is that random.shuffle does the shuffle in place and does not return any output. You can use random.sample instead:

df['new_text'] = df['text'].str.lower().str.split()
df['new_text'] = df['new_text'].apply(lambda x: random.sample(x, k=len(x)))

Resulting values:

0         [is, weather, the, nice]
1        [amazing, house, is, the]
2    [are, blooming, flowers, the]
Name: new_text, dtype: object
  • Related