You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@ignite.apache.org by "Konstantin Orlov (Jira)" <ji...@apache.org> on 2021/01/22 08:32:00 UTC

[jira] [Created] (IGNITE-14034) Calcite integration. indexCondition refactoring

Konstantin Orlov created IGNITE-14034:
-----------------------------------------

             Summary: Calcite integration. indexCondition refactoring
                 Key: IGNITE-14034
                 URL: https://issues.apache.org/jira/browse/IGNITE-14034
             Project: Ignite
          Issue Type: Improvement
          Components: sql
            Reporter: Konstantin Orlov


Currently IndexCondition is quite cumbersome and hard to understand. The difference between bounds and conditions is unclear as well as unclear what should be used to estimate a selectivity and what should be used to estimate a self cost.

Thus I suggest to change it in a follow way:
 * remove [lower|upper]Cond
 * bounds remains as is
 * self cost estimation of an AbstractIndex should be calculated with regard to bounds
 * selectivity should be calculated with regards to whole condition that is member of ProjectableFilterableTableScan



--
This message was sent by Atlassian Jira
(v8.3.4#803005)