Difference in Lookup row count - ODBC and UniVerse Stage
Posted: Tue Apr 24, 2007 2:10 pm
Hi,
Case1 : We had designed a lookup logic where the data for the reference link is fetched using ODBC stage from an Oracle table. We had used ODBC stage because we wanted to make use of the property "Reference link with multi row result set".
Case2 : Then we decided to make use of Hashed File for lookup instead of making multiple database connections which happens in ODBC stage. So, we replaced the ODBC stage with UniVerse stage to read a Hashed file. The Hashed File was written in a separate job. UniVerse stage was used here to read the Hashed File because we wanted to make use of the property "Reference link with multi row result set".
After doing all these, for our surprise, the records returns from the lookup in both these cases were different. In case1, we fetch it directly from the table and in case2, the table is populated into the Hashed file and that is used for the lookup. But how can both these yield different number of records after the lookup.
How can the lookup using OBDC stage and UniVerse stage (Hashed File) give different outcome for the same source? Is there any reason for this?
Thanks.
Case1 : We had designed a lookup logic where the data for the reference link is fetched using ODBC stage from an Oracle table. We had used ODBC stage because we wanted to make use of the property "Reference link with multi row result set".
Case2 : Then we decided to make use of Hashed File for lookup instead of making multiple database connections which happens in ODBC stage. So, we replaced the ODBC stage with UniVerse stage to read a Hashed file. The Hashed File was written in a separate job. UniVerse stage was used here to read the Hashed File because we wanted to make use of the property "Reference link with multi row result set".
After doing all these, for our surprise, the records returns from the lookup in both these cases were different. In case1, we fetch it directly from the table and in case2, the table is populated into the Hashed file and that is used for the lookup. But how can both these yield different number of records after the lookup.
How can the lookup using OBDC stage and UniVerse stage (Hashed File) give different outcome for the same source? Is there any reason for this?
Thanks.