You can do many things .
Index_id : 0 heap, 1 Clustered Index, 2 Nonclustered index
Instead of select count(*) from tb , select sum(rows) from view where table_name = tn and index_id in (0,1)
Selecting from this view is not very fast, as includes many DDViews
Improvement : uses sys.dm_db_partition_stats , that shows how many pages ( 1 page = 8kb) are used or reserved for a given partition
