Fix failures in JPPD tests (#526)

* Fix plan verification failures due to Semi Join

- In case of nested join scenario, multiple RTFs arrives at scan on probe side. This resulted in plan showing multiple RuntimeFilters. This has been fixed and hence changes are required.

- Semi Join is enable by default and the plan has to factor in the changes.

- Disable Semi Join where Agg is required before Join.

* Disable failing queries with a reference to the bug

framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/blocking_op_build_side.sql

@@ -1,6 +1,7 @@
 set `exec.hashjoin.enable.runtime_filter` = true;
 set `exec.hashjoin.runtime_filter.max.waiting.time` = 10000;
 set `planner.enable_broadcast_join` = false;
+set `planner.enable_semijoin` = false;
 
 select
   l.l_shipdate,
@@ -23,4 +24,5 @@ limit 10;
 
 reset `exec.hashjoin.enable.runtime_filter`;
 reset `exec.hashjoin.runtime_filter.max.waiting.time`;
-reset `planner.enable_broadcast_join`;
+reset `planner.enable_broadcast_join`;
+reset `planner.enable_semijoin`;

framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/blocking_op_build_side.sql.explain

@@ -1,6 +1,7 @@
 set `exec.hashjoin.enable.runtime_filter` = true;
 set `exec.hashjoin.runtime_filter.max.waiting.time` = 10000;
 set `planner.enable_broadcast_join` = false;
+set `planner.enable_semijoin` = false;
 
 explain plan for
 select
@@ -24,4 +25,5 @@ limit 10;
 
 reset `exec.hashjoin.enable.runtime_filter`;
 reset `exec.hashjoin.runtime_filter.max.waiting.time`;
-reset `planner.enable_broadcast_join`;
+reset `planner.enable_broadcast_join`;
+reset `planner.enable_semijoin`;

framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/broadcast_hash_01.sql → framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/broadcast_hash_01.sql.fail

@@ -1,3 +1,4 @@
+-- Enable when DRILL-6914 is fixed
 set `exec.hashjoin.enable.runtime_filter` = true;
 set `exec.hashjoin.runtime_filter.max.waiting.time` = 10000;
 set `planner.enable_broadcast_join` = true;

framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/broadcast_hash_02.plan

@@ -17,9 +17,7 @@
 04-03\s+SelectionVectorRemover
 04-04\s+Filter.*
 04-05\s+RuntimeFilter
-04-06\s+SelectionVectorRemover
-04-07\s+RuntimeFilter
-04-08\s+Scan.*lineitem.*columns=\[`l_suppkey`, `l_orderkey`, `l_shipdate`, `l_extendedprice`, `l_discount`\]\]\]\)
+04-06\s+Scan.*lineitem.*columns=\[`l_suppkey`, `l_orderkey`, `l_shipdate`, `l_extendedprice`, `l_discount`\]\]\]\)
 04-02\s+BroadcastExchange
 06-01\s+HashJoin.*
 06-03\s+Scan.*supplier.*columns=\[`s_suppkey`, `s_nationkey`\]\]\]\)

framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/broadcast_hash_02.sql → framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/broadcast_hash_02.sql.fail

@@ -1,3 +1,4 @@
+-- Enable when DRILL-6914 is fixed
 set `exec.hashjoin.enable.runtime_filter` = true;
 set `exec.hashjoin.runtime_filter.max.waiting.time` = 10000;
 set `planner.enable_broadcast_join` = true;

framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/hash_join_date.plan

@@ -6,11 +6,9 @@
 04-02\s+SelectionVectorRemover
 04-03\s+RuntimeFilter
 04-04\s+Scan.*lineitem.*columns=\[`l_shipdate`\]\]\]\)
-01-04\s+HashAgg.*
-01-06\s+Project.*
-01-08\s+HashToRandomExchange.*
+01-04\s+Project.*
+01-06\s+HashToRandomExchange.*
 03-01\s+UnorderedMuxExchange
 05-01\s+Project.*
-05-02\s+HashAgg.*
-05-03\s+Project\(EXPR\$0=\[CAST\(\$0\)\:\DATE\]\)
-05-04\s+Scan.*lineitem.*columns=\[`l_shipdate`\]\]\]\)
+05-02\s+Project\(EXPR\$0=\[CAST\(\$0\)\:\DATE\]\)
+05-03\s+Scan.*lineitem.*columns=\[`l_shipdate`\]\]\]\)

framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/hash_join_date.sql → framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/hash_join_date.sql.fail

@@ -1,3 +1,4 @@
+-- Enable when DRILL-6914 is fixed
 set `exec.hashjoin.enable.runtime_filter` = true;
 set `exec.hashjoin.runtime_filter.max.waiting.time` = 10000;
 set `planner.enable_broadcast_join` = false;

framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/hash_join_double.plan

@@ -6,11 +6,9 @@
 04-02\s+SelectionVectorRemover
 04-03\s+RuntimeFilter
 04-04\s+Scan.*lineitem.*columns=\[`l_discount`\]\]\]\)
-01-04\s+HashAgg.*
-01-06\s+Project.*
-01-08\s+HashToRandomExchange.*
+01-04\s+Project.*
+01-06\s+HashToRandomExchange.*
 03-01\s+UnorderedMuxExchange
 05-01\s+Project.*
-05-02\s+HashAgg.*
-05-03\s+Project\(EXPR\$0=\[CAST\(\$0\)\:DOUBLE\]\)
-05-04\s+Scan.*lineitem.*columns=\[`l_discount`\]\]\]\)
+05-02\s+Project\(EXPR\$0=\[CAST\(\$0\)\:DOUBLE\]\)
+05-03\s+Scan.*lineitem.*columns=\[`l_discount`\]\]\]\)

framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/hash_join_double.sql → framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/hash_join_double.sql.fail

@@ -1,3 +1,4 @@
+-- Enable when DRILL-6914 is fixed
 set `exec.hashjoin.enable.runtime_filter` = true;
 set `exec.hashjoin.runtime_filter.max.waiting.time` = 10000;
 set `planner.enable_broadcast_join` = false;

framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/nested_broadcast.plan

@@ -2,9 +2,7 @@
 03-06\s+HashJoin.*
 03-08\s+SelectionVectorRemover
 03-09\s+RuntimeFilter
-03-10\s+SelectionVectorRemover
-03-11\s+RuntimeFilter
-03-12\s+Scan.*partsupp.*columns=\[`ps_suppkey`, `ps_partkey`\]\]\]\)
+03-10\s+Scan.*partsupp.*columns=\[`ps_suppkey`, `ps_partkey`\]\]\]\)
 03-07\s+BroadcastExchange
 04-01\s+HashJoin.*
 04-03\s+SelectionVectorRemover

framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/nested_broadcast.sql → framework/resources/Advanced/runtimefilter/tpch_sf100/parquet/nested_broadcast.sql.fail

@@ -1,3 +1,4 @@
+-- Enable when DRILL-6792 is fixed
 set `exec.hashjoin.enable.runtime_filter` = true;
 set `exec.hashjoin.runtime_filter.max.waiting.time` = 10000;
 set `planner.enable_broadcast_join` = true;