Low cardinality index

Why we must have an index on type field? As type is a low cardinality field I think whole scan must be faster

Couchbase bucket can have any type of unrelated documents. Normally queries are done on related documents by using type or any others means.
Example: default bucket has
orders 1K
sales 10K
airports 20K

if you want query on orders FULL SCAN does 31K rows, If you do on type you only do 1K

if you want query on orders FULL SCAN does 31K rows, If you do on type you only do 1K

In RDBMS worlds , 31K full scan is faster than 1K index scan + 1K finding data , Why it is not true in couchbase?

@vsr1 reply please …