Identifying indexes for compression

[gfishburn]

I want to know how to identify which indexes would benefit from the Oracle 8i feature of compression.

It will be indexes that have a low cardinality, but how do I identify them?

[gpsingh]

index keys are stored in the form of prefixsufix , if u see the prefix for many keys are same in
a single leaf node, u can choose that index for compression.
,e.g., keys like (city_ny,city_edison,city_jerseycity etc ) in which city_ prefix is common in the keys and hence this key (index column) can be compressed.

But be sure, that this index is not highly update one, else lot of time would be wasted in uncompressing and re -arraning.

You need to know ur data to find out which index can be compressed.


thanks
GP