Splunk Search

How to transform a table

i111040d
New Member

I have this table:

_time,id,src,dst
9:00,x,A,B
9:01,x,B,C
9:02,y,C,B
9:03,z,B,C
9:04,y,B,A
9:05,z,C,D

I wanna create this table:

_time,id,A-B,B-C,C-D
9:00,x,9:00,9:01,
9:02,y,9:04,9:02,
9:03,z,,9:03,9:05

How can I transform the table?

Tags (1)
0 Karma

somesoni2
Revered Legend

Try this

your current search giving _time id src dst | eval temp=src."-".dst."#".strftime(_time,"%H:%M")
| stats first(_time) as _time values(temp) as temp by id 
| mvexpand temp | rex field=temp "(?<direction>.+)#(?<time>.+)"
| eval temp=_time."#".id | chart values(time) over temp by direction | rex field=temp "(?<_time>.+)#(?<id>.+)" | fields - temp

Update#1
Thanks @woodcock for clarifying the requirement. This should take care of normalization as well.

 your current search giving _time id src dst | eval Nodes=mvsort(split(src."-".dst,"-")) | eval temp=mvindex(Nodes,0)."-".mvindex(Nodes,1)."#"._time | stats first(_time) as _time values(temp) as temp by id 
 | mvexpand temp | rex field=temp "(?<direction>.+)#(?<time>.+)"  | eval temp=_time."#".id | chart values(time) over temp by direction | rex field=temp "(?<_time>.+)#(?<id>.+)" | fields - temp| table _time id *
0 Karma

woodcock
Esteemed Legend

Yours does not normalize B-A to be A-B so the results are wrong:

_time   id  A-B B-A B-C C-B C-D
9:00    x   9:00        9:01         
9:02    y       9:04        9:02     
9:03    z           9:03        9:05

Also, you must change strftime(_time, "%H,%M") to just _time. Use this to spoof test events:

|noop|stats count AS trash|eval trash="9:00,x,A,B::9:01,x,B,C::9:02,y,C,B::9:03,z,B,C::9:04,y,B,A::9:05,z,C,D"
|makemv delim="::" trash|mvexpand trash
| rex field=trash "(?<_time>.*?),(?<id>.*?),(?<src>.*?),(?<dst>.*)" | table _time id src dst
| rename DesiredResults AS "
_time,id,A-B,B-C,C-D
9:00,x,9:00,9:01,
9:02,y,9:04,9:02,
9:03,z,,9:03,9:05"

somesoni2
Revered Legend

Not sure if that was the requirement (normalization). Would agree on strftime change if event though the field name is _time, use is storing string formatted time in it.

0 Karma

woodcock
Esteemed Legend

Normalization is specified in the desired results given.

0 Karma

somesoni2
Revered Legend

I did miss that. The answer updated to take care of both.

0 Karma

woodcock
Esteemed Legend

Add this:

| rex field=_time mode=sed "s/://"
| eval transition = src . "-" . dst . "::" . dst . "-" . src | fields - src dst
| makemv delim="::" transition | mvexpand transition
| stats values(transition) AS transition BY _time id | eval transition = mvindex(transition, 0)
| chart first(_time) AS _time BY id transition
| eval time="9999" | foreach * [ eval time = min(time, $<<FIELD>>$) | rex field="<<FIELD>>" mode=sed "s/(\d{1,2})(\d{2})/\1:\2/" ]
| rex field=time mode=sed "s/(\d{1,2})(\d{2})/\1:\2/"
| table time id *
0 Karma
Got questions? Get answers!

Join the Splunk Community Slack to learn, troubleshoot, and make connections with fellow Splunk practitioners in real time!

Meet up IRL or virtually!

Join Splunk User Groups to connect and learn in-person by region or remotely by topic or industry.

Get Updates on the Splunk Community!

Announcing Modern Navigation: A New Era of Splunk User Experience

We are excited to introduce the Modern Navigation feature in the Splunk Platform, available to both cloud and ...

Observability Simplified: Combining User Experience, Application Performance & ...

Tech Talk Observability Simplified: Combining User Experience, Application Performance & Network ...

Event Series May & June: From Network Visibility to Service Intelligence

Unifying the Network: Moving from Alert Noise to Service Intelligence with Splunk ITSI In today’s hybrid ...