How to get transaction results in table with "transacted" items in separate rows?

[splunk_thunk]

Hello Experts,

I have a transaction query that I am displaying in a table. I am able to get results in a table, however, the results tied in a single transaction appear as single row in the table. I would like to have them displayed in separate rows, as if they are individual search results.

Here's an example:

Log Data:

  Transaction Id=1, step=1, data_x=dataX1, data_y=dataY1

  Transaction Id=1, step=2, data_x=dataX2, data_y=dataY2

How results look like

Transaction Id	data_x	data_y
1 1	dataX1 dataX2	dataY1 dataY2
2 2	...	...

Need it to look like

Transaction Id	data_x	data_y
1	dataX1	dataY1
1	dataX2	dataY2
2	...	...
2	...	...

Any help appreciated.

Thanks!

[richgalloway]

Please share the SPL that produces the current results.

[splunk_thunk]

@richgalloway 

| makeresults count=4
| streamstats count
| eval transactionId = case(count=1 OR count=2, 1, count=3 OR count=4, 2) | eval step = case(transactionId=1 AND count<=2,count, transactionId=2,count-2) | eval data_x="dataX".step | eval data_y="dataY".step | eval val=case(step=1,step,step>1,"null. So how to print value_1 here as well?") | transaction transactionId | table transactionId, step, val, data_x, data_y | sort transactionId

Please try the above.

I have another question (you will notice it in search results from above query). For the same transaction, how can I print the value of a field from step=1 of the transaction into other steps in the result, when that field is not available in the other search results within same transaction.

Thanks much!

[ITWhisperer]

I assume you used transaction to gather the events together so you could try using stats and list()

The main thing you need to do is zip all the fields together, mvexpand, then "unzip" them. This does rely on there being values in each of the fields, although that would be true for transaction as well

| makeresults count=4
| streamstats count
| eval transactionId = case(count=1 OR count=2, 1, count=3 OR count=4, 2) | eval step = case(transactionId=1 AND count<=2,count, transactionId=2,count-2) | eval data_x="dataX".step | eval data_y="dataY".step | eval val=case(step=1,step,step>1,"null. So how to print value_1 here as well?")
| stats list(data_*) as data_* list(step) as step list(val) as val by transactionId
| eval data=mvzip(data_x,mvzip(data_y,val,":"),":")
| mvexpand data
| eval data_x=mvindex(split(data,":"),0)
| eval data_y=mvindex(split(data,":"),1)
| eval val=mvindex(split(data,":"),2)
| table transactionId data_x data_y val

[splunk_thunk]

Thanks @ITWhisperer . That solves one problem. 

Any suggestions to copy value from step-1 of the transaction into step-2 of the same transaction when it is missing in step-2?

I have added  this table to a dashboard that has filter on val from my example. When val field is not available in step-2 of the same transaction, the filter condition removes that result from the table, which is not desired. 

[ITWhisperer]

| makeresults count=4
| streamstats count
| eval transactionId = case(count=1 OR count=2, 1, count=3 OR count=4, 2) | eval step = case(transactionId=1 AND count<=2,count, transactionId=2,count-2) | eval data_x="dataX".step | eval data_y="dataY".step | eval val=case(step=1,step,step>1,null())
| streamstats latest(val) as val by transactionId
| stats list(data_*) as data_* list(step) as step list(val) as val by transactionId
| eval data=mvzip(data_x,mvzip(data_y,val,":"),":")
| mvexpand data
| eval data_x=mvindex(split(data,":"),0)
| eval data_y=mvindex(split(data,":"),1)
| eval val=mvindex(split(data,":"),2)
| table transactionId data_x data_y val

[splunk_thunk · Accepted solution]

Thanks @ITWhisperer . Trying it out in prod before I mark resolved.