Understanding ROWID in Data Distribution

TimesTen Scaleout requires a unique id for row distribution. It uses ROWID to ensure uniqueness across all elements.

For tables with a duplicate distribution scheme where K-safety is set to 1 and for all tables (no matter what the distribution scheme is) where K-safety is set to 2, the physical location of each copy of a row is different, so each copy of the row has different ROWID values. In this case, when using ROWID based access, TimesTen Scaleout returns the value of the ROWID in the first data space. If the row in the first data space is not available, TimesTen Scaleout returns the ROWID in the next (second) data space.

Since ROWID is the identifier of a specific copy of a row, if that copy is not available, you cannot access the row by ROWID. In this case, you should access the row by primary key.

See ROWID Pseudocolumn in Oracle TimesTen In-Memory Database SQL Reference.


Applications should not store ROWID values in the database and try to use these values later. Applications can fetch the ROWID in a transaction and then use the ROWID later in the same transaction.