tree a28719e110f2ec6ab9dd2693233e1c0fe650737a
parent e7fa4b3fb1662f1ae6d1a2336e2be6b51e18f22d
author Taewoo Kim <wangsaeu@yahoo.com> 1539046883 -0700
committer Taewoo Kim <wangsaeu@gmail.com> 1539133127 -0700

[NO ISSUE][COMP][RT] Enable multiway similarity joins

- Enable the FuzzyJoinRule that transforms
  a nested-loop-similarity-join plan to a three-stage-similarity join.
- Modify FuzzyJoinRuleCollections.
  - Add the ExtractCommonExpressionRule to extract common expressions
    in the star-like multiple similarity join substitutions.
  - Add the InlineSubplanInputForNestedTupleSourceRule to translate
    the generated subplan from the similarity function-derived
    substitution into join in case of nested schemas.
  - Use similarity-jaccard-prefix to enable the pp+ join strategy.
  - Use the right side to build the heavy hash join on
    the prefix tokens from both sides.
  - Add RemoveAssign/Variables/AggRules to iteratively remove unused
    assign/vars once FuzzyJoinRule is applied in each round.
- Add three new optimization cases for multiway similarity joins.
  - link-like multiway similarity joins
  - star-like multiway similarity joins
  - hybrid multiway similarity joins with the both styles of similarity joins.
- Add a check whether a similarity function is on
  a select over an existing similarity join.
- Change the inverted-index-based similarity join to the three-stage-similarity join
  due to efficiency considerations.

Change-Id: I8736f104905eeda763d39709e002c2b9629278cc
Reviewed-on: https://asterix-gerrit.ics.uci.edu/1076
Sonar-Qube: Jenkins <jenkins@fulliautomatix.ics.uci.edu>
Tested-by: Jenkins <jenkins@fulliautomatix.ics.uci.edu>
Contrib: Jenkins <jenkins@fulliautomatix.ics.uci.edu>
Integration-Tests: Jenkins <jenkins@fulliautomatix.ics.uci.edu>
Reviewed-by: Dmitry Lychagin <dmitry.lychagin@couchbase.com>
Reviewed-by: Taewoo Kim <wangsaeu@gmail.com>