sonveer .If the file records are physically ordered on a non-key field such as name (which does not have a distinct value for each record) then in that case clustered index is used.
And the index is maintained for every distinct value of that clustering field. So it is not purely dense or purely sparse, right?
for every distinct value
In Primary Indexing, the Primary Key(P.K.) is used as anchor attribute. It is an example of sparse indexing.
If the number of blocks acquired by index file is n, then Block Access required will be = log2n + 1.
In Clustered Indexing, there will be one entry for each unique value of the non-key attribute.
If the number of blocks acquired by index file is n, then Block Access required will be >= log2n + 1.
*differences are many, this is the major one to note.