How to deal with CSV log file with missing end timestamps

steadyfish

Hi, I'm a beginner in using ProM tool. I want to know if there's an optimal way around treating events with missing end timestamps. 

Based on various tutorials and forum discussions  (https://svn.win.tue.nl/trac/prom/export/22567/Documentation/LogCSVImport.pdf), I've prepared my event log. In this specific scenario, I have a few events with start and end timestamps; a few others only have a start timestamp.

On importing this csv log into ProM, I get errors - "Error: java.text.ParseException: Could not parse NULL timestamp!" (Sending NULL values as blanks resulted in the same error).

One solution to this could be setting end timestamps same as start timestamps when they are missing, but I wanted to understand if there's a better way to do this.

sample log below - 

--Case-No-------|--Event--------|--Timestamp--Start---|--Timestamp--End-----|

--1234----------|--A-----------|--2016-01-01 10:00:00|--NULL---------------|

--1234----------|--B-----------|--2016-01-01 10:00:00|--2016-01-01 11:00:00|

--1234----------|--C-----------|--2016-01-01 10:00:00|--NULL---------------|

Thanks,
Dhrumin S

JBuijs

Dear Dhrumin,

Thank you for your question.
My suggestion would be to set the error handling setting (last wizard screen) to 'skip errors' (or similar) such that you just don't get the completion timestamps.
If they are missing, using the start time will mess with your performance analysis. Especially if the end timestamps are missing for particular activities always this is the best approach.

Happy mining!