Hi All,
I'm trying to combine a number of fields using:
| stats values(task_name) as task_name by idnumber
This works great when it comes to timestamps associated with the idnumber, but for
the tasks associated with it, splunk sorts it alphabetically.
This leads to problems down the line when we try to see which task was executed first.
Part of the problem is that the number of timestamps can differ from the number of tasks
so to make a new field with timestamp and task combined does not work.
#original data:
sysmodtime,task_name,idnumber
05/01/20 12:00 PM,one,1
05/01/20 12:01 AM,two,1
05/01/20 12:02 AM,two,1
05/01/20 12:02 AM,two,1
05/02/20 12:00 PM,one,2
04/02/20 12:00 AM,one,2
04/02/20 01:00 AM,one,2
04/02/20 02:00 AM,one,3
05/04/20 12:00 PM,one,4
05/03/20 12:00 PM,two,4
05/03/20 12:01 PM,three,4
05/03/20 12:02 PM,four,4
05/03/20 12:40 PM,five,4
05/03/20 12:50 PM,six,4
#the conflicting results after stats command (see attachment)
Any advice would be welcome
Cheers,
Roelof
It's not clear what results you want. Can you provide a mockup of the desired output?
Have you tried putting both columns in the stats command?
| stats values(*) as * by idnumber
Hi Richgalloway, thanks for the reply, the atachment in the main question shows the erroneous results, with the tasks sorted alphabetically instead of matched with the timestamp, so for example for id number 1 there are 6 tasks and 6 time stamps, the stats command (placing both columns inside stats) gives:
sysmodtime, task_name, idnumber
05/04/20 12:00 PM, five, 4
05/03/20 12:00 PM, four,
05/03/20 12:01 PM, one,
05/03/20 12:02 PM, six.
05/03/20 12:40 PM, three,
05/03/20 12:50 PM, two,
whereas the order should be simply; one, two, three, four, five, six
Solutions that make a new field using sysmodtime and task_name fail since number of tasks
and number of sysmodtime are not equal in all cases.