Question

I am wondering if a table or particularly a heap file with a clustered index created on one of the attributes is sorted on that attribute.Is there a difference between a sorted heap file and a heap file with clustered index ?

Answer 1

Is there a difference between a sorted heap file and a heap file with clustered index ?

Sorted heap file is a heap (array-like structure) with the elements ordered. This is what MyISAM does with ALTER TABLE … ORDER BY … or PostgreSQL does with CLUSTER. Inserting a new record into such a table breaks the order.

Clustered index is a B-Tree. Inserting a new record keeps the order.

The records in a clustered table are not necessarily physically ordered (in a sense that the record with a higher value of the clustering key has a higher offset in a file or tablespace).

Rather, the records are logically ordered: you can traverse from one record to another following B-Tree links (or direct links between pages in case of a B+Tree), but this still implies random disk seeks because the pages linked logically do not necessarily reside near each other on the disk.

Some databases, Oracle for instance, allow reading the indexes in physical order (using an operation known as INDEX FAST FULL SCAN) which does not maintain the logical order but is faster because of the benefits or sequential access to the pages.

There is no such thing as a “heap file with a clustered index”: a table organization is either heap or clustered.