spark-instrumented-optimizer/sql/catalyst
Dongjoon Hyun c632bdc01f [SPARK-14029][SQL] Improve BooleanSimplification optimization by implementing Not canonicalization.
## What changes were proposed in this pull request?

Currently, **BooleanSimplification** optimization can handle the following cases.
* a && (!a || b ) ==> a && b
* a && (b || !a ) ==> a && b

However, it can not handle the followings cases since those equations fail at the comparisons between their canonicalized forms.
* a < 1 && (!(a < 1) || b)     ==> (a < 1) && b
* a <= 1 && (!(a <= 1) || b) ==> (a <= 1) && b
* a > 1 && (!(a > 1) || b)     ==> (a > 1) && b
* a >= 1 && (!(a >= 1) || b) ==> (a >= 1) && b

This PR implements the above cases and also the followings, too.
* a < 1 && ((a >= 1) || b )   ==> (a < 1) && b
* a <= 1 && ((a > 1) || b )   ==> (a <= 1) && b
* a > 1 && ((a <= 1) || b)  ==> (a > 1) && b
* a >= 1 && ((a < 1) || b)  ==> (a >= 1) && b

## How was this patch tested?

Pass the Jenkins tests including new test cases in BooleanSimplicationSuite.

Author: Dongjoon Hyun <dongjoon@apache.org>

Closes #11851 from dongjoon-hyun/SPARK-14029.
2016-03-22 10:17:08 -07:00
..
src [SPARK-14029][SQL] Improve BooleanSimplification optimization by implementing Not canonicalization. 2016-03-22 10:17:08 -07:00
pom.xml [SPARK-11565] Replace deprecated DigestUtils.shaHex call 2016-02-10 09:52:35 +00:00