CODEWITHSUNDEEP
pythonpandasseries
Shew
Shew

Reputation: 1596

Unique value count in pandas series

I have a panda series of long strings.

I want to get the value counts of words in the entire series. I tried with 

df.value_counts().to_dict()

But it gives the string level count not the word level count.

How can I do it efficiently ?.

My series looks like below

print df.head(3)

0 4632 N. Rockwell Street, Chicago Rockwell Neighborhood 773 60625 4748 N Kedzie

1 4632 N. Rockwell Street, Chicago Rockwell' Bdoy 773 60625 4632 N Rock

2 4632 N. Rockwell Street, LA Rock hood Grill 773 60625 3658 W Lawren

I wanted to generate a dictionary as follows

a['4632'] = 3
a['Rockwell'] = 3
a['LA'] = 1

and so on

Upvotes: 2

Views: 220

Answers (1)

jezrael
jezrael

Reputation: 862431

I think here is better pure python solution with Counter of joined all values to long string with split:

from collections import Counter

d = Counter(' '.join(df).split())
#if necessary convert to dict
#d = dict(Counter(' '.join(df).split()))

Or use split with stack first:

d = df.str.split(expand=True).stack().value_counts().to_dict()
print (d)
{'Rockwell': 4, '4632': 4, 'Street,': 3, '773': 3, '60625': 3, 'N.': 3, 'N': 2, 'Rock': 2, 'Chicago': 2, 'Kedzie': 1, 'Grill': 1, 'Neighborhood': 1, '3658': 1, 'Lawren': 1, 'W': 1, '4748': 1, 'LA': 1, 'hood': 1, "Rockwell'": 1, 'Bdoy': 1}

Upvotes: 2

Related Questions