CODEWITHSUNDEEP
pythonpandasdataframe
Oeyvind
Oeyvind

Reputation: 357

Count occurrences in DataFrame across columns

I have a table like this:

     Actor1 Actor2 Actor3
 0   Bob    Alice  Tom
 1   Frank  Bob    Sam

I want to count the number of times each actor occurs anywhere. In other words, Bob = 2, while the others are 1.

I know how to count per column and add, and I know I can merge each column into a very long Series. But there must be a better way?

Upvotes: 1

Views: 130

Answers (1)

busybear
busybear

Reputation: 10590

You could use np.unique and set return_counts=True. This will return two lists with the unique values and counts.

np.unique(df, return_counts=True)

Output:

(array(['Alice', 'Bob', 'Frank', 'Sam', 'Tom'], dtype=object),
 array([1, 2, 1, 1, 1], dtype=int64))

Otherwise you can use df.stack().value_counts() to which you (and @ShubhamSharma!) allude.

Upvotes: 3

Related Questions