Count occurences of a field value depending on another field

[airmouli]

Hello,

I have a set of data similar to this :

session1 | user1 | computer 1 | start
session2 | user2 | computer 2 | start
session1 | user1 | computer 1 | stop
session2 | user2 | computer 2 | stop
session1 | user1 | computer 1 | start
session3 | user3 | computer 3 | start

I would like to count the number of starts and stops for each session.

session1 | user1 | computer 1 |    2     |       1
session2 | user2 | computer 2 |    1     |       1
session3 | user3 | computer 3 |    1     |       0

Thank you for your help

[kmorris_splunk · Accepted solution]

Try something like this:

[YOUR BASE SEARCH HERE] 
| stats count(eval(action="start")) as starts count(eval(action="stop")) as stops by session user computer

I created a field called action for start and stop values, as well as giving the other fields logical names: session; user; and computer.

[kmorris_splunk]

I just realized my search was a bit off based on your request. You wanted to know the number of starts and stops per session. That would look more like the following:

[YOUR BASE SEARCH HERE] 
 | stats count(eval(action="start")) as starts count(eval(action="stop")) as stops values(user) as users values(computer) as computers by session

The users and computers fields would have a multivalued list of all distinct values for the user and computer fields. Not sure if that is what you wanted, but probably a good idea since there would be multiple values based on your data sample.

[airmouli]

That's exactly what I was looking for ; Thank you very much