Suggesting to use full table scan rather than going with any kind of Index(cluster/non-cluster Index) seek/scan in following context:
- While accessing all the records of your table. Based on your comment - "all the record need to be fetched every-time", so it's not effective to access data via any kind of index.
- Table rows contains 10-20K rows and updated frequently. Based on your comment - "...have 10334 records in it now, not expected to grow beyond 15K. The table is frequently updated."
In addition, cluster index doesn't maintain data structure(table) separately rather it sorts and stores the data rows in the table itself.
Answer for your Questions:
Q# But will it be a good idea, keeping table as a heap?
Answer# There is no harm if you keep table as a heap(without cluster index) and access all rows all the time without ordering by specific Coloumn for the result set.
Q# As all the record need to be fetched every-time, will a table scan and Clustered-index scan give same performance?.
Answer# Get Clustered-index if you fetch records in a sorted order at any point of time.