Our customer has a log file with fix-width layout as below,
10/4/12 18:00:00 10/4/12 18:00:31 ABCDEFGHIJKLM- Completed
NOPQ
10/4/12 18:00:01 10/4/12 18:00:22 1234567890123- Completed
456
It expect a field, let's say "name", be extracted with value "ABCDEFGHIJKLNOPQ" and "1234567890123456". How do we use transform or anything else to parse it?
Are they two different events?
You should be able to use something like the following to test the extraction on your data: (example using rex from search commands and piping to stats for demo)
* | rex field=_raw "\d+\:\d+\:\d+\s+(?P<name>\w+)\-\s+" | stats count by name
The easiest way for you to set up the extraction is using the IFX (Interactive Field eXtractor), please see follow:
http://docs.splunk.com/Documentation/Splunk/5.0/Knowledge/ExtractfieldsinteractivelywithIFX
Or you can create field extractions in conf files, please see following:
Thanks for your answer. The event contains two lines and the field "ABCDEFGHIJKLM-" and "NOPQ" are located at separated two lines.
So there are two questions in this case exactly. First, how to extract column field from multiple lines? Second, how to tell between hyphen and dash?