Hi,

When running a query with Order By clause using JdbcAdapter, I observed the
JdbcSort rule is not being used.

I have try a few variation of order by and the calcite always fall to
using EnumerableSort.

I tried to workaround by it override JdbcSort.computeSelfCost to return a
very minimal value (CPU = 1, ROW = 1, IO  = 0)

but the cheapest plan always fall to using EnumerableSort even though the
EnumerableSort.computeSelfCost return a higher value. (CPU:100.0 io:0.0
row:3684.1361487904733)

Can someone give me some pointer why this does not work and how the
cheapest plan identification work?

I have also tried override getRows and that does not work neither.


Thank you.

Rgds,
jay

Reply via email to