SQL Power Business Intelligence Productivity Tools
Company OverviewBusiness Intelligence Productivity ToolsQuick-Start Implementation ServicesDemos & TurotialsFrequently Asked Questions (FAQ)Open Source Community ResourcesSQL Power ForumImplementation & Technology PartnersGet SQL Power SoftwareContact Us

SQL Power Software Forum

SQL Power Software Forum

  [Search] Search   [Recent Topics] Recent Topics   [Hottest Topics] Hottest Topics   [Members]  Member Listing   [Groups] Back to home page  [Register] Register /  [Login] Login 
Messages posted by: lilihammer
Forum Index » Profile for lilihammer » Messages posted by lilihammer
Author Message
speaking with the users, a 'null' value in a lighter colour would be sufficient.
I think your hypothesis was correct. I created a new project and there was no longer an error.

Maybe that could be a suggestion/enhancement for the product, to have the ability to adapt to a table changing the number of columns it has.
OS: Windows
Version of MatchMaker: 0.9.3-alpha

Situation Causing Error:
Created a De-duping project and ran the engine with no errors encountered.
Running the match engine again after adding a filter to the project.
The following error message appeared.

Error Message:
java.lang.RuntimeException: java.lang.IndexOutOfBoundsException: Index: 29, Size: 29
at ca.sqlpower.matchmaker.MatchEngineImpl.call(Unknown Source)
at ca.sqlpower.matchmaker.swingui.engine.EngineWorker.doStuff(Unknown Source)
at ca.sqlpower.swingui.SPSwingWorker.run(SPSwingWorker.java:73)
at java.lang.Thread.run(Unknown Source)
Caused by: java.lang.IndexOutOfBoundsException: Index: 29, Size: 29
at java.util.ArrayList.RangeCheck(Unknown Source)
at java.util.ArrayList.get(Unknown Source)
at org.hibernate.collection.PersistentList.get(PersistentList.java:249)
at ca.sqlpower.matchmaker.munge.SQLInputStep.doCall(Unknown Source)
at ca.sqlpower.matchmaker.munge.AbstractMungeStep.call(Unknown Source)
at ca.sqlpower.matchmaker.munge.MungeProcessor.call(Unknown Source)
at ca.sqlpower.matchmaker.MatchEngineImpl.mungeAndMatch(Unknown Source)
... 4 more
Rumor has it that you should be able to apply a filter on munge processes. Is this true?

In the 0.9.3-alpha version for windows, there is no such option.
It would be a great option to have.
In the Validate Matches window of the Match Maker show null values as an empty cell.

Showing this to the user may be a faster indicator of what values are empty.
just a suggestion, arranging the munge processes in order of their priority level.
i think currently the munge processes are arranged in the tree according to the order they were made in.
As part of the "Validate Matches" screen it may be pertinent to also have the number of records that ran through the MatchMaker engine. I'm not sure, maybe this was intended to be a part of the Match Statistics that is not yet available?
Along the same lines as being able to select which munge processes to run. Sometimes the user only wants to validate matches based on specific munge processes. To have the ability to select which ones to view for validation purposes would be amazing.
Actually, I thought that a straight export of the actual screen for "Validation Status" would be sufficient.

When creating munge processes, give the user the ability to preview some of the data and see what data transformations are occurring. This would help the user when they are developing their munge processes; to ensure they are matching based on what they think they are matching on.

For example when creating a munge process such as the one in the attached file. The user would be able to click on the Retain Characters plug in and preview the transformed input data.
There should be some sort of acknowledgment that the engine is doing something after the user clicks on a button if there is a delay to the action.

For example: Running the engine, the user clicks on Abort!, there is a delay. The message indicating the stop of the engine, takes some time to appear, so to inform the user that the engine is doing something pop up a note saying "Abort in Process" or have the cursor turn into an hour glass.
It would be useful for testing and debugging to be able to select certain munge processes in a given project to be selected and ran.

For example: If a project contains 20 different munge processes, maybe the user would only like to run 2 of them in the first run. Then on the second run the user would then like to run 5 munge processes.
OS: Windows
Version of MatchMaker: 0.9.3-alpha

Issue: In the Run Match Engine window, after running the match engine, clicking on the Match Statistics... button yields an error message indicating that match statistics is not yet available. But when clicking on Match statistics from the tree the match statistics can be viewed.
You would expect that you could also view the match statistics from clicking on the button in the run match engine window.
OS: Windows
Version of MatchMaker: 0.9.3-alpha

Issue:
Created a De-duping project with 8 munge processes.
Right clicked on the project and selected Duplicate Project.
Accepted the default naming and the location the Project was saved.
The duplicate project was present but the duplicate munge processes did not recognize the source table and the connections from the input to the munge processes and results were gone (see attached jpeg). The munge processes could not be run without building them again.

Expected:
The duplicated project to be at a state where the engine could run it.


Currently: You can now export the Munge pen to a PDF file (although, I've never tried this)

Suggestion: Give the ability to the user to be able to export the List of Munge Processes, and Validation Statistics to a PDF file.
 
Forum Index » Profile for lilihammer » Messages posted by lilihammer
Go to:   
Powered by JForum 2.1.8 © JForum Team