topic Re: How to extract data from log message data using rex field=_raw? in Splunk Search

How to extract data from log message data using rex field=_raw?

bharath_hk12 — Thu, 07 Dec 2023 08:01:28 GMT

Hi,

I have logger statements like below:

Event data - {"firstName":"John","lastName":"Doe"}

My query needs <rex-statement> where double quotes (") in the logs are parsed and the two fields are extracted in a table:

index=my-index "Event data -" | rex <rex-statement> | fields firstName, lastName | table firstName, lastName

Please let me know what <rex-statement> do I have to put. Thanks in advance.

Re: How to extract data from log message data using rex field=_raw?

gcusello — Thu, 07 Dec 2023 08:11:50 GMT

Hi @bharath_hk12 ,

at first, this seems to be a json log, so you can use the spath command (https://docs.splunk.com/Documentation/Splunk/9.1.2/SearchReference/Spath).

Anyway, you could use a regex like the following:

| rex "\"firstName\":\"(?<firstname>[^\"]+)\".\"lastName\":\"(?<lastName>[^\"]+)"

that you can test at https://regex101.com/r/XwqxZB/1

to be more sure, you should share some complete samples of your logs (not partial).

Ciao.

Giuseppe

Re: How to extract data from log message data using rex field=_raw?

bharath_hk12 — Thu, 07 Dec 2023 08:33:36 GMT

Thanks for your response @gcusello. Could you suggest how to use spath for the same json log? How to extract json out of the log statement and further individual fields.

I am not able to share the actual logs, since they are client specific.

Re: How to extract data from log message data using rex field=_raw?

bharath_hk12 — Thu, 07 Dec 2023 08:35:30 GMT

There are some 15-20 fields within JSON, but I want to extract only 3-4 of them. So if spath serves the purpose, good for me

Re: How to extract data from log message data using rex field=_raw?

inventsekar — Thu, 07 Dec 2023 10:16:22 GMT

>>> I am not able to share the actual logs, since they are client specific.

to use the spath command, we should check the json format.

so, pls remove all client specific things like ip address, hostnames, usernames, etc.. replace them with sample data... like firstname:adam, etc.. and paste the json..

Re: How to extract data from log message data using rex field=_raw?

bharath_hk12 — Thu, 07 Dec 2023 10:30:50 GMT

I have already done that in the main question.

Here's the sample, just that the fields will be more:

Event data - {"firstName":"John","lastName":"Doe"}

Re: How to extract data from log message data using rex field=_raw?

yuanliu — Thu, 07 Dec 2023 17:52:57 GMT

Do you mean to say that the raw event data contains extra strings such as "Event data -"? Is there any additional strings after the closing curly bracket? Is there any other opening curly bracket before the one shown in your mock data? If none of those exist, simply do

| rex "(?<json>{.+)" | spath input=json

Here is an emulation using your mock data:

| makeresults | eval _raw = "Event data - {\"firstName\":\"John\",\"lastName\":\"Doe\"}" | rex "(?<json>{.+)" | spath input=json | fields - _*

Play with it and compare with real data. The emulation should give

firstName	json	lastName
John	{"firstName":"John","lastName":"Doe"}	Doe

Re: How to extract data from log message data using rex field=_raw?

bharath_hk12 — Fri, 08 Dec 2023 12:15:58 GMT

Thanks @yuanliu, I was able to use this to derive from my actual logs