Splunk Search

Conditional count distinct if

jclemons7
Path Finder

Hello all,

I'm looking to do a "count distinct value if record type = foobar" type of scenario. Hopefully, I'll be able to articulate what I'm trying to do here.

record type A:

record: person
name: bob
id: 123456
sex: m
state: tx
hp: 555-123-1234
dept: finance



record: person
name: jane
id: 7949191
sex: f
state: ca
hp: 555-456-7890
dept: marketing



record: person
name: jane
id: 7949191
sex: f
state: ca
hp: 555-456-7890
dept: marketing

record type B:

record: computer
computername: mycomputer
type: pc
ram: 4GB
ip: 1.1.1.1
dept: finance





record: computer
computername: mylaptop
type: laptop
ram: 4GB
ip: 2.2.2.2
dept: finance





record: computer
computername: theserver
type: server
ram: 16GB
ip: 10.0.0.1
dept: marketing

stats if(record="computer",dc(computername)) as dc_computer if(record="person",dc(id)) as dc_people by dept

desired result:

dept        dc_computer dc_people   
finance         2           1                       
marketing         1         2                       

somesoni2
Revered Legend

Give this a try. There is no id field for computer and there is no computername field for person record, it can be simply done by this

your base search | stats dc(computername) as dc_computer dc(id) as dc_people by dept 
0 Karma

jclemons7
Path Finder

ugh.. sorry, in my "real" data.. the field I need to count distinct values from is actually the same name. I think my dummy data is not accurate..

0 Karma

sideview
SplunkTrust
SplunkTrust

I like this answer but it does rely on the additional assumption that the computername is also unique. iow there might be some "my laptop" computers out there. or just null values which wouldn't get counted.

0 Karma

sideview
SplunkTrust
SplunkTrust

I recommend breaking this sort of thing into an eval statement and then a separate stats command just to make it more comprehensible.

Also in this case you can make a separate "record_type" field and then use that with the chart command to do it a little more compactly.

| eval record_type=case(record="computer","computer",  record="person","person")
| chart dc(id) over department by record_type

However, there is another way, and this I think is more the way you were headed -- to create separate id fields and then use stats

| eval person_id=if(record="person",id,null()
| eval computer_id=if(record="computer",id,null())
| stats dc(person_id) dc(computer_id) over department 
Get Updates on the Splunk Community!

Stay Connected: Your Guide to May Tech Talks, Office Hours, and Webinars!

Take a look below to explore our upcoming Community Office Hours, Tech Talks, and Webinars this month. This ...

They're back! Join the SplunkTrust and MVP at .conf24

With our highly anticipated annual conference, .conf, comes the fez-wearers you can trust! The SplunkTrust, as ...

Enterprise Security Content Update (ESCU) | New Releases

Last month, the Splunk Threat Research Team had two releases of new security content via the Enterprise ...