K
Keith Sudbury
Guest
For a pretty wild guess you can estimate based on the records per block, number of rows to delete, average row size and a fudge factor of how well packed the data you plan to delete is concentrated per DB block. To get the exact numbers try the delete against a copy of the database 
(with a dbanalys before and after). This has the advantage of knowing the new free blocks, an approximate length of time for the deletes and a safe way to check to make sure you only delete what you wanted to. Depending on the percentage of data you are removing per table you might be better off with copying the data you want to keep to an identical table (different name obv) and then renaming the tables afterwards.
Continue reading...
(with a dbanalys before and after). This has the advantage of knowing the new free blocks, an approximate length of time for the deletes and a safe way to check to make sure you only delete what you wanted to. Depending on the percentage of data you are removing per table you might be better off with copying the data you want to keep to an identical table (different name obv) and then renaming the tables afterwards.Continue reading...