My company had a partitioned view across 5 SQL 2000 and 2005 clusters, with 1.9B rows of data, with the key using datatype of int. At this point, we need to convert everything to bigint very soon. We were hoping to change the view to cover both int and bigint so that we do not have to change all 1.9B rows of data to bigint. However, we are running into performance issue with these views.
Here is the setup:
- Clusters of 2000 and 2005, joined by partitioned view
- PK on int on old tables, with check constraint on each table
- New tables set up with bigint data type
- All check constraints converting the data types to bigint
Any suggestion would be appreciated.