How to count the number of upper case words more than 1 character long in series

Question

I have a series of strings and I'm trying to create a new column that counts the number of upper case words in each string, with the constraint that the word is greater than 1. For example, the series

s = pd.Series(['I AM MAD!', 'Today is a nice day', 'This restaurant SUCKS'])

would return a series with values of 2, 0, 1.

A few other helpful questions on here have shown me one way to do this for a single string:

sum(map(str.isupper, [word for word in s[0].split() if len(word) > 1]))

which correctly returns 2.

But I'm wondering how to apply this to the entire series without looping over each element?

Answer 1

You can use regex to extract the words, and then count:

(s.str.extractall(r'(\b[A-Z]{2,}\b)')  # extract all capitalized words with len at least 2
  .groupby(level=0).size()             # count by each index
  .reindex(s.index, fill_value=0)      # fill the missing with 0
)

Output:

0    2
1    0
2    1
dtype: int64

Answer 2

Borrow Quang's regex

s.str.count(r'(\b[A-Z]{2,}\b)')
0    2
1    0
2    1
dtype: int64