Splunk Search

Join not working

kavyamohan
Explorer
index="event"
| rex field=Macaddress mode=sed "s/(.{2})/\1-/g s/-$//"
 | rename Macaddress as "macAddress" 
| eval SessionTime_epoch = strptime(SessionTime, "%Y-%m-%dT%H:%M:%S.%3NZ")
 | eval SessionTime = strftime(SessionTime_epoch, "%Y-%m-%d %H:%M:%S")
| join macAddress type=outer
    [inputlookup nctdata.csv]
| rename cubicleNo as "work_station" | join work_station type=outer
    [inputlookup somaster.csv]
|table Username,macAddress,Session,SessionTime,SystemName,cubliceNo,vlanNo, c_occupiedclient

My query is something like this. But i am not getting the results correctly. I want the fields as I have mentioned above. I want Username,macAddress,Session,SessionTime,SystemName from "event" file and join macAddress in nctdata and get these columns in nctdata file: cubliceNo,vlanNo and join work_station in nctdata and somaster file and take these column in somaster:c_occupiedclient.

Tags (2)
0 Karma
1 Solution

gcusello
SplunkTrust
SplunkTrust

Hi kavyamohan,
the join command is a very slow solution and has the limit of 50,000 results because there's a subsearch.
In your case you want to do a join with a lookup, to do this you don't need of join and you can use the lookup command that's like a join.
Don't think to Splunk as a DB, it's different!

Instead, try to use a different approach:

index="event"
| rex field=Macaddress mode=sed "s/(.{2})/\1-/g s/-$//"
| rename Macaddress as "macAddress" 
| eval SessionTime_epoch = strptime(SessionTime, "%Y-%m-%dT%H:%M:%S.%3NZ"), SessionTime = strftime(SessionTime_epoch, "%Y-%m-%d %H:%M:%S")
| lookup nctdata.csv macAddress OUTPUT cubicleNo 
| rename cubicleNo as "work_station" 
| lookup somaster.csv work_station OUTPUT cubliceNo vlanNo
| table Username macAddress Session SessionTime SystemName cubliceNo vlanNo c_occupiedclient

In addition, when you use the inputlookup command you have to always use the pipe char "|".

Bye.
Giuseppe

View solution in original post

gcusello
SplunkTrust
SplunkTrust

Hi kavyamohan,
the join command is a very slow solution and has the limit of 50,000 results because there's a subsearch.
In your case you want to do a join with a lookup, to do this you don't need of join and you can use the lookup command that's like a join.
Don't think to Splunk as a DB, it's different!

Instead, try to use a different approach:

index="event"
| rex field=Macaddress mode=sed "s/(.{2})/\1-/g s/-$//"
| rename Macaddress as "macAddress" 
| eval SessionTime_epoch = strptime(SessionTime, "%Y-%m-%dT%H:%M:%S.%3NZ"), SessionTime = strftime(SessionTime_epoch, "%Y-%m-%d %H:%M:%S")
| lookup nctdata.csv macAddress OUTPUT cubicleNo 
| rename cubicleNo as "work_station" 
| lookup somaster.csv work_station OUTPUT cubliceNo vlanNo
| table Username macAddress Session SessionTime SystemName cubliceNo vlanNo c_occupiedclient

In addition, when you use the inputlookup command you have to always use the pipe char "|".

Bye.
Giuseppe

wmyersas
Builder

You can even simplify the eval further:

| eval SessionTime = (strftime(strptime(SessionTime, "%Y-%m-%dT%H:%M:%S.%3NZ"), "%Y-%m-%d %H:%M:%S")
0 Karma

kavyamohan
Explorer

Thank you so much it worked:)

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!

SOK it to Me: Top 3 Benefits of Using Splunk Operator on Kubernetes that’ll Make ...

    Thursday, July 9, 2026  |  11:00AM–12:00PM PDT Duration: 1 hour (includes Q&A) Managing can feel like a ...

Upgrade Prep for 10.4, Network Observability Deep Dives, and More from Splunk Lantern

Splunk Lantern is Splunk’s customer success center that provides practical guidance from Splunk experts on key ...

Splunk Developer Day announcements: AI agents, MCP tools, Forecasting, and Custom ...

Splunk Developer Day was packed with product and platform updates for developers building in the AI ...