DevOps Articles

Attribute Clustering for Performance

by One of the nice facilities in 12c is the ability to instruct the database to co-locate data during loading in order to facilitate reduced resource consumption for subsequent queries.

So let’s now populate (create) our table T with a straight copy of the data from SOURCE_DATA and index a column of interest You can see the data is in pseudo-random order on both OBJECT_ID and OWNER.

Just under 5400 consistent gets, which is not dissimilar from the rows we fetched because those rows from OBJECT_ID 10,000 to 15,000 they are scattered throughout the table.