To prevent spam users, you can only post on this forum after registration, which is by invitation. If you want to post on the forum, please send me a mail (h DOT m DOT w DOT verbeek AT tue DOT nl) and I'll send you an invitation in return for an account.

Key/Value Set plugin

bartsmeets
edited December 2011 in ProM 6
Hello,

I finally managed to get Prom6 running (it appartently had something to do with my Java version), but now I have been trouble importing my log. I have a log like this:

Patient ID | Activity | Timestamp
100 | act1 | 01-12-2011 08:00
100 | act2 | 01-12-2011 08:30
101 | act1 | 01-12-2011 08:00

When I import this file(CSV) in Prom6 it is shown as a key/value set. I then use the 'Convert Key/Value Set to Log' plugin in order to create a log. To trace identifier I assign the PatientID column, to concept:name the Activity column and to time:timestamp the Timestamp column. The problem is that this never gives the desired result. Most of the time it correctly identifies the 'instances' (the PatientID) and the event related to them (Activity) but it does not recognize the timestamp ( - no timestamp - ). In rare cases it does recognize the timestamp, but then it does not connect the activity names to the events. In other cases it thinks the activities are instances... This happens seemingly arbitrarily but with exact the same mapping setup... So far it has been a pain to get the correct log from my CSV file. Does anyone have a solution for this?

Alternatively I could convert it using XESame, however this does not seem to run correctly. Even when I set it up as noted in the tutorial it will always give the following error: "Sorry but the test failed because the driver could not be found: sun.jdbc.odbc.JdbcOdbcDriver".

It would help me out a great deal if I could simply convert my CVS to a usable log.

Thanks in advance!

Bart

Answers

  • Hi Bart,

    Regarding your first problem, can you make sure that the CSV is correctly formatted? Could you for instance try to load it in Excel and check if all the columns are correctly matched? It could be that you have an extra or missing " somewhere that incorrectly starts a new column...

    Regarding the XESame problem, what JRE/Java version are you using? If you're not using the official JRE than this explains the error and you need to find the corresponding JDBC-ODBC driver class for that JRE.

    Please let me know what works or what not.
    Joos Buijs

    Senior Data Scientist and process mining expert at APG (Dutch pension fund executor).
    Previously Assistant Professor in Process Mining at Eindhoven University of Technology
  • I've managed to get it working. Apparently when I connect the timestamp (in the csv) to the timestamp and the lifecycle attribute it always assign a timestamp to the activities in the resulting log. Not sure why this works, but it does :)

    Thanks for the support!
  • Hi, I have the same problem. When converting from a key/value set to a log, no timestamp is taken although specified in the csv-file. Taking the lifecycle attribute instead works, but then I have a NullPointerException when I want to add an artificial start/end event (because of the missing timestamp). I also tried running-example.csv of the Process Mining book, but also in this case the timestamp is not taken. Is there a special format that must be considered?

Sign In or Register to comment.