Compute and broadcast a count in pandas (with groupby transform)

Question

How can I compute and broadcast a count in pandas?

To compute a count:

df.groupby('field').size()

To broadcast an aggregation to the original dataframe:

df.groupby('field')['field_to_aggregate'].transform(aggregation)

The latter works if I specify the field to aggregate onto and aggregations like sum, mean, etc. But I am not finding a way to make it work when I want a simple count of the grouped-by field.

(Note: I could just use the first and re-join on the original table against the grouped-by table, but I want to avoid joins and I'm looking for an efficient solution that uses pandas' transform)

Answer 1

You could try:

result = df.groupby('field')['field_to_aggregate'].transform('size')

Note that 'field_to_aggregate' can be the same as 'field'.