Open Forum

Expand all | Collapse all

MDA and Integration Manager for Payables

  • 1.  MDA and Integration Manager for Payables

    Posted Feb 05, 2019 03:55 PM


    I'm implementing MDA for use primarily with AP, but I'm running into issues with setting up the integration.  I can get the GL Distributions integrated successfully but the MDA details will only work if each GL Distribution account on the AP transaction is unique. If a transaction has two separate GL Distributions to the same account, the amount for that distribution gets applied to every MDA Group/Code that shares that same account in the transaction. Any ideas on how to fix this?  Also, is it possible to include transactions with and without MDA in a single integration?  Every time I try, I get a "system out of memory" error in IM.


    Gina Behee
    Director of Accounting
    Pharmacy First
    Shawnee Mission KS

  • 2.  RE: MDA and Integration Manager for Payables

    Posted Feb 06, 2019 11:05 AM
    ​I'm not an expert on multi-dimensional analysis, but from what it sounds like you are running into a problem that DBA's refer to as a "Cartesian join".  That happens when you are expecting unique rows but don't actually get unique rows and so as the data is being matched up it is getting multiplied into MANY more rows.  To illustrate the number of rows, you take the number of records from one data source and MULTIPLY it times the number of records in your second datasource.  Add a third and you can see where things quickly get out of hand.

    My suspicion is that in order for MDA to work properly, the accounts MUST be unique.  I think that this is where you are getting duplication of data and the subsequent performance issues.  And my bet is that there really isn't any way around this constraint.

    Blair Christensen
    Database Administrator
    Oppenheimer Companies, Inc.
    Boise ID

  • 3.  RE: MDA and Integration Manager for Payables

    Posted Feb 07, 2019 12:06 PM

    ​I have done integrations of journal entries with MDA data included, but it isn't pretty. I started with an existing sample integration that uses 4 separate source files. There is no sorting on any of the files. The important thing is to have a "sequence number" column as the unique identifier across the files. I have copied the relationship screen below, as well as each data file and file properties. In this example, I have MDA codes on each transaction, but I have also done integrations with codes on only some of the transactions; in that case, those lines would not be in the "group" and "code" files. Though this is for journal entries, I would imagine that the AP import would be similar

    I hope this helps.

    File 1 - Date

    File 2 - Detail of all transactions

    File 3 - MDA groups

    File 4 - MDA codes

    Sandy Righi
    Senior Programmer Analyst
    Howard Hanna
    Mayfield Heights OH

If you've found this thread useful, dive deeper into User Group community content by role