Hi,
As heavily skewed table decreases performance, planning to recreate skewed tables automatically with a new Primary index. Plan is to have an automated script which would fetch all the skewed tables (> 40%) owned by my team and analyze on the best index and recreate the table.
Difficult part seems to be automatically predicting suitable primary index. [Like, in case of summary tables - have to neglect metric columns from index and only choose the dimensions].
Hoping some of you would have faced similar situation before. Kindly help / share info on script to automatically choose PI for a table (interms of distribution & not having to add all the columns as PI).
Regards,
Saran
Forums: