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.

[ProM-developers] Conversion to mxml

JBuijsJBuijs Posts: 910
edited December 2010 in - ProMimport

On Friday 5-11-2010, 10:03 CET the following message was posted on the prom developers mailinglist:

Good Morning Doctors,

 

I am an italian student of Politecnico of Torino
(Management Engineering).

I need your help for a conversion of a file Access to a
log file for ProM 5.2.

First, let me explain why : I am doing a relation about
your software ProM and the ideas and purposes inside the software (why exists
and why you create it...) ; a main part of this relation, which is in fact a
short thesis, consists in an application of the software with a simple example
and a simple database: a comparison between results with proM and other way
results.

So i need the software for a statistical analysis to
prove the accuracy of your software.

 

I downloaded your proM import framework and i tried to
convert my file, but my knowledge of Java 'language' is very bad and i can't
solve errors.

To go on, i need that someone could convert this file for
me.

 

I ask you if it is possible to have a converted and
tested version of the file, please.

 

I leave the file attached to this mail, if you gently
could help me.

 

Best Wishes,

 

Niccolò Milano

Post edited by JBuijs on
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

Comments

  • JBuijsJBuijs Posts: 910
    To which I replied at 10:24:

    Good Morning Niccolò,

     

    Unfortunately, we cannot convert your database for you
    since converting data already requires some decisions to be made which can
    influence the process mining results. So, converting the data is also part of
    the process mining process.

     

    There are two options for you if you want to convert your
    data to MXML:

    1) Use ProM Import. In this case the best option would be
    to follow the 'general purpose ProMImport plug-in' tutorial which can be found
    at http://prom.win.tue.nl/research/wiki/promimport/tutorials.
    This would require you to create 4 tables and copy the data to these 4 tables
    in a special way (explained on the website).

    2) Another option would be to use XESame. This tool
    allows you to define trace and event extractions on databases. Having glanced
    at your database structure I think this will be the best option for you. XESame
    can be found at http://prom.win.tue.nl/research/wiki/xesame/start
    and has the option to export to XES (used in ProM 6) but also MXML.

     

    Furthermore, I'm personally interested in your report. If
    you could send me a draft/concept/final version of your thesis when ready I
    would be very interested.

     

    Kind regards,

     

    J.C.A.M. (Joos) Buijs MSc

    PhD student

    Fac. Wiskunde & Informatica

     

    Den Dolech 2, 5612 AZ

    P.O. Box 513, HG 7.43

    5600 MB Eindhoven

    The Netherlands

     

    T +31 40 247 3661

    M +31 6 42 77 89 88

    F +31 40 246 3992

    j.c.a.m.buijs@tue.nl

    http://www.tue.nl/staff/j.c.a.m.buijs



    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
  • JBuijsJBuijs Posts: 910
    Niccolo Milano replied on 15-11:

    Good Evening Doctor Buijs,







    Thank you very much for your help and support two weeks
    ago.
    Unfortunately i had lots of work to do, so i could work
    on my conversion only few days ago.
    I tried the two options but with no positive results:
    1) in this case, i built the 4 tables but a error's
    message compairs:

    ERROR: ImportFilter 'MS Access database' has died during
    import routine

    ERROR: Error message was:

    ERROR: Fatal error in class 'MS Access database'

    Exception: java.sql.SQLException:[Microsoft][Driver ODBC
    Microsoft Access] impossible to find the file 'unknown'

    Original stack trace:

    createSQLException in sun.jdbc.odbc.JdbcOdbc
    (null,line-1) standardError in ...

    SQLDriverConnect in ...

    ...

    (I've already followed the tutorial to configure
    correctly the odbc connection to MsAccess).

    2)if it easier comparing to the first option, how could i
    load a mapping file? i mean, how could i create a mapping file from my excel
    file? i tried using the csv format with no success.

    In the file attached there are the 4 tables in access, so
    maybe you can have a look and tell me what's wrong with it.

    Thanks for your help and your patience,

    Kind Regards,

    Niccolò Milano

    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
  • JBuijsJBuijs Posts: 910
    Dear Niccolo,

    First of all, I'm not a doctor (yet), I'm still a PhD student :)
    Second, sorry for my late reply but I was out of office for 2 days.

    As to your problems:
    1) Did you test your ODBC connection to your MS Access database? And did you close your MS Access database when you ran ProM Import? The error seems to indicate that it can either not find the specified file or that it is locked.
    Where you able to successfully complete the ProM Import tutorial???

    2) In my opinion (but I'm biased since I programmed XESame) using XESame is easier than manually filling the ProM Import MS Access database and then converting it using ProM Import (option 1).
    The first step you need to make is to create an ODBC connection to your source file (which is your Excel file I suppose). In Windows there are ODBC drivers for this. Another option would be to create a .csv file for each workspace in your Excel file, place these .csv files in one directory and then create an csv file ODBC connection to this directory. Again, Windows contains such an ODBC driver.
    The next step would be to download ProM 6 which includes XESame. Then start XESame, connect to your ODBC connection you just made and enter the mapping.
    This is described in my Master's Thesis of which a link can be found at the XESame Process Mining Wiki Page.

    I hope this gets you going. Let me know if you encounter any more problems.

    Kind regards,

    Joos
    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
  • JBuijsJBuijs Posts: 910
    On 19-11 Niccolo replied:

    Dear Joos,

    1)i solved one error closing the file before launch it, but it didn't still work. I will try again configuring the odbc database in windows, even if i completed successfully the tutorial (and the conversion of the file
    runningexample...)

    so 2) i am reading your detailed Master's Thesis to solve my problems with xesame. Today i will try to convert the file.

    Thank you again for your support,

    I will let you know if i'll still have problems.

    Kind Regards,

    Niccolò Milano
    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
  • JBuijsJBuijs Posts: 910
    And on 23-11 he reported:

    Dear Joos,

    I read you thesis; of course I skipped some parts that don't need to my conversion (e.g. case study 1).
    The test connection works. So i moved to the 'mapping' part. My case casually is on your thesis (''If one patient is considered to be the process instance one can investigate which treatments patients undergo including possible relationships between different treatments'').

    Now, I have just an excel paper with this tables: idpatient, idtreatment, nametreatment, datetreatment, typerecovery.

    So, if I am right, I added only a new event called Exams.

    I put the name of the excel file in the property FROM, then the values in the attributes.

    But..when i try to convert, ERROR: at the step 2of5 compares a message telling that there is an error in the table from and the Microsoft Jet module manager can't read the object, be sure that exists.
     
    in the case study you put the excel file (in csv) in the FROM...
    so i tried also converting first the file in csv, do again the odbc driver mode in windows, then the test connection (passed), then put in the from the file in csv...
    but nothing happened..

    If it could help you, i send my mapping file and my excel file, so you can see what's wrong (i hope not everything...)

    Kind Regards,

    Niccolò Milano
    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
  • JBuijsJBuijs Posts: 910
    To which my reply was on 24-11:

    Dear Niccolo,

    First a general remark: there is a ProM forum where you can also post your questions. The benefit is that there are more people that can read and reply to your questions as it might be that I'm not able to respond for a couple of days. Furthermore, this serves as a sort of archive for questions that people ask and their answers.
    If you post anything on the forum, I will be notified and reply as quickly as I would to an e-mail.
    I have created a discussion from your e-mails which I will update sometime soon. The direct link is: http://prom.win.tue.nl/forum/index.php?p=/discussion/comment/89/#Comment_89

    As to your question: whether the table name refers to the Excel file name or the worksheet name within that file depends on the Excel driver that you use. See http://support.microsoft.com/kb/178717
    Could you read that webpage, try to correct your query and report back if it works? Unfortunately, today I'm too busy to look in details into your problem.
    If you are still stuck, let me know so I can help you in more detail in a couple of days.

    Kind regards,

    Joos
    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
Sign In or Register to comment.