We are using iDM 764 and UniVerse adapter to connect to UniVerse database. I generated a synonym for a table which has multivalued fields. When I retrieve the sample data for the synonym I am not seeing any data for the mutlivalued columns. Is there any configuration in the master file/synonym I need to do to extract multivalue field data?

EDASPROF has the following-
ENGINE SQLUV SET CONNECTION_ATTRIBUTES UnivDev...
ENGINE SQLUV SET OPENMODE NNF

According to the R764 doco:
The OPENMODE NNF setting determines how tables with multivalued columns are treated.
For the Adapter for Universe to function properly, OPENMODE must be set to NNF in
edasprof.prf.

Since you already have that, have you tried creating a simple procedure to read the table rather than using the 'Sample Data' option? With some types of data I have seen the Sample Data option not work.

Using "Sample Data" wizard and a procedure which has a UV SQL I am able to fetch the data in all the columns in the table except the multivalued columns.

Is there a technique to fetch the data in multi-valued columns other than setting the OPENMODE to NNF?

I'm using the default settings for the UniVerse connector. And I'm always getting all multivalued fields. Even the CO fields are created in masterfiles.

Within UniVerse you can setup witch table's and columns are accesible. You should check this.

Nice to see someone else out there that uses a multivalue database.

Does anyone know if this adapter works with Raining Data D3?

If it does is there documentation on this especially about defining multivalues?

How do you define the pick master dictionary to define the file?

This message has been edited. Last edited by: jammer,

I set something up years ago, and read multi values. From memory it reads them as records, similar to an OCCURS in a master.

I got the UniVerse adapter configuration to work. I came to know that the UniVerse DBA has setup the multi-valued column in table#1 as a seperate table so that this data can be accessed by ODBC clients. UniVerse adapter uses the ODBC client to talk to the database. Once I knew the name of this table, I was able to create a synonym and fetch the data.

One lesson learned here is UniVerse supports wild characters like "*", "-" in their table names. So if a table in UniVerse is called "Customer-Data*345", the ODBC layer translates it to "Customer_Data_345". So when creating a synonym we have to use to ODBC name not the native UniVerse table name.

The name of the table doesn't depend on Universe, it depends on how the dictionary's (DICT) in Universe are created by your DBA.

The special charachters are removed because it causes problems in SQL statements.

One thing to keep in mind: you can't sort on multivalued fields.