If you need the access count of indexes (hash indexes or join indexes), then you are referring to the correct view.
Access of primary index and secondary index can be different.
I guess that you are referring to the count of the columns accessed. That can be fetched using 'DBC.ColumnsVX' view
PS: If the count is not getting logged, make sure you have enabled dbscontrol flag ''ObjectUseCountCollectRate'
Note that the count measured is the count from the 'Optimizer execution plan'.
(Count for explain, insert explain, dump explain not accounted)
you can also use DBQLobjects for that purpose. There is ObjectType field to filter the index rows there.
It would be interesting to compare numbers from dbc.qrylogobjects and from dbc.indicesv for your example.
With kind regards,
were you able to turn on the DBQLobjects and see the results?
It would not answer the question why counters for different columns are different. But would answer the initial question about how many times the index is used.
Also note that DBQL provides more detailed information than AccessCounts. You would know what users and queries are using that index.
Look at the DBQLobjTBL or DBQLOBJTBL_HST. There is a TypeOfUse column. With values like this
The values can also be summed up, if a column is encountered in more than one condition. Such as