How can I retrieve count or distinct count of some field values using stats function

[phaniraj]

I have a table in this form (fields and values):

  USERID    USERNAME     CLIENT_A_ID      CLIENT_B_ID
  11        Tom          555              123
  11        Tom          555              456
  11        Tom          777              456
  11        Tom          999              456
  22        Jill         444              789
  22        Jill         888              234

The output of the splunk query should give me:

  USERID    USERNAME     CLIENT_A_ID_COUNT      CLIENT_B_ID_COUNT
  11        Tom          3                      2
  22        Jill         2                      2

Should calculate distinct counts for fields CLIENT_A_ID and CLIENT_B_ID on a per user basis.

[round_the_twist]

For second part of question , 

use stats dc(client_a_id), dc(client_b_id) by USERNAME

[daryllj]

Thanks for this as well- very helpful (and worked).

[Stephen_Sorkin]

Use:

... | stats dc(CLIENT_*_ID) as CLIENT_*_ID_COUNT by USERID USERNAME

[HXCaine]

Might be worth using 'distinct_count' rather than 'dc' on these boards, as it wasn't obvious to me what 'dc' was until I discovered distinct_count

[daryllj]

I just checked this one out, and seems to be a quick way of getting what I need:

index="google" sourcetype="*directory*" "emails{}.address"="*mydomain.com"
| timechart dc(name.fullName) span=1d cont=FALSE

Thanks for the pointer- I have not used the dc command before.