Grouping by two fields, want to get distinct count of values in second field

Hi, I wrote the following Splunk query which returns a list of distinct USER_AGENTs for each SESSION_ID: index=abc | rex field=_raw "-S:(?\w+)-.+User agent: '(?.+)', Referrer" | stats count by SESSION_ID, USER_AGENT I would now like to modify this query to return a list of SESSION_IDs that have more than one unique value for USER_AGENT, and the count of the unique values. Thanks! Jonathan