Extract a field using Regex

[sanjay_shrestha]

Hi,

I have following output from a log file.

 

(5/1/13 - 1:36:05.01 PM)    Event   LOAD 1 Setup

(5/1/13 - 1:36:08.01 PM)    Event   LOAD 2 Setup

(5/1/13 - 1:37:07.37 PM)    Event   LOAD 1 Process

(5/1/13 - 1:37:17.37 PM)    Event   LOAD 3 Process

(5/1/13 - 1:38:07.39 PM)    Event   LOAD 1 Complete

(5/1/13 - 1:38:15.01 PM)    Event   LOAD 3 Setup

(5/1/13 - 1:38:17.39 PM)    Event   LOAD 2 Complete

(5/1/13 - 1:39:07.42 PM)    Event   READ 1 Setup

(5/1/13 - 1:39:17.37 PM)    Event   LOAD 3 Process

(5/1/13 - 1:39:27.39 PM)    Event   LOAD 3 Complete

(5/1/13 - 1:39:37.42 PM)    Event   READ 2 Setup

(5/1/13 - 1:39:57.42 PM)    Event   READ 3 Setup

(5/1/13 - 1:40:07.45 PM)    Info    READ 1 Process

(5/1/13 - 1:41:07.47 PM)    Error   READ 1 Complete

(5/1/13 - 1:41:17.45 PM)    Info    READ 2 Process

(5/1/13 - 1:41:27.45 PM)    Info    READ 3 Process

(5/1/13 - 1:41:57.47 PM)    Error   READ 2 Complete

(5/1/13 - 1:42:07.47 PM)    Error   READ 3 Complete

I need to extract a field "WorkID", so I used following REGEX

 rex field=_raw "LOAD (?\d+)|READ (?\d+)"

and got error message "Regex: two named subpatterns have the same name"

If I change the WorkID field to WorkID1 and WorkID2, it works but not sure how to consolidate these 2 fields.

Later I will be using "Transaction" to get following output:

Start Time                    End Time                       WorkId 

(5/1/13 - 1:36:05.01 PM)    (5/1/13 - 1:41:07.47 PM)                1

(5/1/13 - 1:36:08.01 PM)    (5/1/13 - 1:41:57.47 PM)                2

(5/1/13 - 1:38:15.01 PM)    (5/1/13 - 1:42:07.47 PM)                3

What would be best ( practice) implementation for this issue?

Thanks!!!!

[sanjay_shrestha]

Thanks to wpreston and sdaniels. Both solutions worked!!!!

[wpreston · Accepted solution]

If you rework your RegEx a little bit, you should be able to get the field extraction. Try this out:

rex field=_raw "(LOAD|READ)\s(?<workid>\d+)"

[sdaniels]

I think this comes down to a better regex. This is an example that will match just the workid number you are looking for or you could change your 'OR' on the word match as well. A lookbehind assertion on a 4 letter word that is all caps plus the space.

(?<=[A-Z]{4}\s)\d

Rex would be something like this:

rex field=_raw "(?<=[A-Z]{4}\s)(?<workid>\d+)"