"Event log mining (text mining)"

p3kka

Hi,

Is there a way, first to tokenize an event log to rows and then filter (with regexp) the time from the single row to an attribute? So the result should have all the rows one by one and each row should have an attribute which contains the timestamp

One event looks basically like this: "Component <time> Event". These kind of events can be thousands in a one log file and multiple log files should be able to handle at one time. Each event is in its own row.

I have achieved the tokenization to rows but how the get all the timestamps to each row for some kind of attribute that is usable afterwards? I'm using the Text Processing expansion.

-p3kka