fix issue 558
diff --git a/asterix-algebra/src/main/java/edu/uci/ics/asterix/optimizer/base/RuleCollections.java b/asterix-algebra/src/main/java/edu/uci/ics/asterix/optimizer/base/RuleCollections.java
index 70ed7ef..a24a459 100644
--- a/asterix-algebra/src/main/java/edu/uci/ics/asterix/optimizer/base/RuleCollections.java
+++ b/asterix-algebra/src/main/java/edu/uci/ics/asterix/optimizer/base/RuleCollections.java
@@ -63,6 +63,7 @@
import edu.uci.ics.hyracks.algebricks.rewriter.rules.ComplexUnnestToProductRule;
import edu.uci.ics.hyracks.algebricks.rewriter.rules.ConsolidateAssignsRule;
import edu.uci.ics.hyracks.algebricks.rewriter.rules.ConsolidateSelectsRule;
+import edu.uci.ics.hyracks.algebricks.rewriter.rules.EliminateGroupByEmptyKeyRule;
import edu.uci.ics.hyracks.algebricks.rewriter.rules.EliminateSubplanRule;
import edu.uci.ics.hyracks.algebricks.rewriter.rules.EnforceOrderByAfterSubplan;
import edu.uci.ics.hyracks.algebricks.rewriter.rules.EnforceStructuralPropertiesRule;
@@ -82,6 +83,7 @@
import edu.uci.ics.hyracks.algebricks.rewriter.rules.IntroduceGroupByForSubplanRule;
import edu.uci.ics.hyracks.algebricks.rewriter.rules.IntroduceProjectsRule;
import edu.uci.ics.hyracks.algebricks.rewriter.rules.IsolateHyracksOperatorsRule;
+import edu.uci.ics.hyracks.algebricks.rewriter.rules.LeftOuterJoinToInnerJoinRule;
import edu.uci.ics.hyracks.algebricks.rewriter.rules.PullSelectOutOfEqJoin;
import edu.uci.ics.hyracks.algebricks.rewriter.rules.PushAssignBelowUnionAllRule;
import edu.uci.ics.hyracks.algebricks.rewriter.rules.PushAssignDownThroughProductRule;
@@ -162,6 +164,8 @@
condPushDownAndJoinInference.add(new PushProperJoinThroughProduct());
condPushDownAndJoinInference.add(new PushGroupByThroughProduct());
condPushDownAndJoinInference.add(new NestGroupByRule());
+ condPushDownAndJoinInference.add(new EliminateGroupByEmptyKeyRule());
+ condPushDownAndJoinInference.add(new LeftOuterJoinToInnerJoinRule());
return condPushDownAndJoinInference;
}
diff --git a/asterix-app/src/test/resources/optimizerts/queries/q08_group_by.aql b/asterix-app/src/test/resources/optimizerts/queries/q08_group_by.aql
new file mode 100644
index 0000000..96b20a4
--- /dev/null
+++ b/asterix-app/src/test/resources/optimizerts/queries/q08_group_by.aql
@@ -0,0 +1,149 @@
+drop dataverse q08_group_by if exists;
+
+create dataverse q08_group_by;
+
+use dataverse q08_group_by;
+
+create type LineItemType as closed {
+ l_orderkey: int32,
+ l_partkey: int32,
+ l_suppkey: int32,
+ l_linenumber: int32,
+ l_quantity: double,
+ l_extendedprice: double,
+ l_discount: double,
+ l_tax: double,
+ l_returnflag: string,
+ l_linestatus: string,
+ l_shipdate: string,
+ l_commitdate: string,
+ l_receiptdate: string,
+ l_shipinstruct: string,
+ l_shipmode: string,
+ l_comment: string
+}
+
+create type OrderType as closed {
+ o_orderkey: int32,
+ o_custkey: int32,
+ o_orderstatus: string,
+ o_totalprice: double,
+ o_orderdate: string,
+ o_orderpriority: string,
+ o_clerk: string,
+ o_shippriority: int32,
+ o_comment: string
+}
+
+create type CustomerType as closed {
+ c_custkey: int32,
+ c_name: string,
+ c_address: string,
+ c_nationkey: int32,
+ c_phone: string,
+ c_acctbal: double,
+ c_mktsegment: string,
+ c_comment: string
+}
+
+create type SupplierType as closed {
+ s_suppkey: int32,
+ s_name: string,
+ s_address: string,
+ s_nationkey: int32,
+ s_phone: string,
+ s_acctbal: double,
+ s_comment: string
+}
+
+create type NationType as closed {
+ n_nationkey: int32,
+ n_name: string,
+ n_regionkey: int32,
+ n_comment: string
+}
+
+create type RegionType as closed {
+ r_regionkey: int32,
+ r_name: string,
+ r_comment: string
+}
+
+create type PartType as closed {
+ p_partkey: int32,
+ p_name: string,
+ p_mfgr: string,
+ p_brand: string,
+ p_type: string,
+ p_size: int32,
+ p_container: string,
+ p_retailprice: double,
+ p_comment: string
+}
+
+create dataset LineItem(LineItemType)
+ primary key l_orderkey, l_linenumber;
+create dataset Orders(OrderType)
+ primary key o_orderkey;
+create dataset Customer(CustomerType)
+ primary key c_custkey;
+create dataset Supplier(SupplierType)
+ primary key s_suppkey;
+create dataset Nation(NationType)
+ primary key n_nationkey;
+create dataset Region(RegionType)
+ primary key r_regionkey;
+create dataset Part(PartType)
+ primary key p_partkey;
+
+for $s in dataset("Supplier")
+ for $lnrcop in (
+ for $lnrco in (
+ for $l in dataset('LineItem')
+ for $nrco in (
+ for $o in dataset('Orders')
+ for $nrc in (
+ for $c in dataset('Customer')
+ for $nr in (
+ for $n1 in dataset('Nation')
+ for $r1 in dataset('Region')
+ where $n1.n_regionkey = $r1.r_regionkey and $r1.r_name = 'AMERICA'
+ return { "n_nationkey": $n1.n_nationkey }
+ )
+ where $c.c_nationkey = $nr.n_nationkey
+ return { "c_custkey": $c.c_custkey }
+ )
+ where $nrc.c_custkey = $o.o_custkey
+ return {
+ "o_orderdate" : $o.o_orderdate,
+ "o_orderkey": $o.o_orderkey
+ }
+ )
+ where $l.l_orderkey = $nrco.o_orderkey
+ and $nrco.o_orderdate >= '1995-01-01'
+ and $nrco.o_orderdate <= '1996-12-31'
+ return {
+ "o_orderdate": $nrco.o_orderdate,
+ "l_partkey": $l.l_partkey,
+ "l_discount": $l.l_discount,
+ "l_extendedprice": $l.l_extendedprice,
+ "l_suppkey": $l.l_suppkey
+ }
+ )
+ for $p in dataset("Part")
+ where $p.p_partkey = $lnrco.l_partkey and $p.p_type = 'ECONOMY ANODIZED STEEL'
+ return {
+ "o_orderdate": $lnrco.o_orderdate,
+ "l_discount": $lnrco.l_discount,
+ "l_extendedprice": $lnrco.l_extendedprice,
+ "l_suppkey": $lnrco.l_suppkey
+ }
+ )
+ where $s.s_suppkey = $lnrcop.l_suppkey
+ return {
+ "o_orderdate": $lnrcop.o_orderdate,
+ "l_discount": $lnrcop.l_discount,
+ "l_extendedprice": $lnrcop.l_extendedprice,
+ "l_suppkey": $lnrcop.l_suppkey,
+ "s_nationkey": $s.s_nationkey
+ }
\ No newline at end of file
diff --git a/asterix-app/src/test/resources/optimizerts/queries/q09_group_by.aql b/asterix-app/src/test/resources/optimizerts/queries/q09_group_by.aql
new file mode 100644
index 0000000..3c9dc47
--- /dev/null
+++ b/asterix-app/src/test/resources/optimizerts/queries/q09_group_by.aql
@@ -0,0 +1,153 @@
+drop dataverse q09_group_by if exists;
+
+create dataverse q09_group_by;
+
+use dataverse q09_group_by;
+
+create type LineItemType as closed {
+ l_orderkey: int32,
+ l_partkey: int32,
+ l_suppkey: int32,
+ l_linenumber: int32,
+ l_quantity: int32,
+ l_extendedprice: double,
+ l_discount: double,
+ l_tax: double,
+ l_returnflag: string,
+ l_linestatus: string,
+ l_shipdate: string,
+ l_commitdate: string,
+ l_receiptdate: string,
+ l_shipinstruct: string,
+ l_shipmode: string,
+ l_comment: string
+}
+
+create type OrderType as closed {
+ o_orderkey: int32,
+ o_custkey: int32,
+ o_orderstatus: string,
+ o_totalprice: double,
+ o_orderdate: string,
+ o_orderpriority: string,
+ o_clerk: string,
+ o_shippriority: int32,
+ o_comment: string
+}
+
+create type CustomerType as closed {
+ c_custkey: int32,
+ c_name: string,
+ c_address: string,
+ c_nationkey: int32,
+ c_phone: string,
+ c_acctbal: double,
+ c_mktsegment: string,
+ c_comment: string
+}
+
+create type SupplierType as closed {
+ s_suppkey: int32,
+ s_name: string,
+ s_address: string,
+ s_nationkey: int32,
+ s_phone: string,
+ s_acctbal: double,
+ s_comment: string
+}
+
+create type NationType as closed {
+ n_nationkey: int32,
+ n_name: string,
+ n_regionkey: int32,
+ n_comment: string
+}
+
+create type RegionType as closed {
+ r_regionkey: int32,
+ r_name: string,
+ r_comment: string
+}
+
+create type PartType as closed {
+ p_partkey: int32,
+ p_name: string,
+ p_mfgr: string,
+ p_brand: string,
+ p_type: string,
+ p_size: int32,
+ p_container: string,
+ p_retailprice: double,
+ p_comment: string
+}
+
+create type PartSuppType as closed {
+ ps_partkey: int32,
+ ps_suppkey: int32,
+ ps_availqty: int32,
+ ps_supplycost: double,
+ ps_comment: string
+}
+
+create dataset LineItem(LineItemType)
+ primary key l_orderkey, l_linenumber;
+create dataset Orders(OrderType)
+ primary key o_orderkey;
+create dataset Supplier(SupplierType)
+ primary key s_suppkey;
+create dataset Region(RegionType)
+ primary key r_regionkey;
+create dataset Nation(NationType)
+ primary key n_nationkey;
+create dataset Part(PartType)
+ primary key p_partkey;
+create dataset Partsupp(PartSuppType)
+ primary key ps_partkey, ps_suppkey;
+create dataset Customer(CustomerType)
+ primary key c_custkey;
+
+for $p in dataset('Part')
+ for $l2 in (
+ for $ps in dataset('Partsupp')
+ for $l1 in (
+ for $s1 in (
+ for $s in dataset('Supplier')
+ for $n in dataset('Nation')
+ where $n.n_nationkey = $s.s_nationkey
+ return {
+ "s_suppkey": $s.s_suppkey,
+ "n_name": $n.n_name
+ }
+ )
+ for $l in dataset('LineItem')
+ where $s1.s_suppkey = $l.l_suppkey
+ return {
+ "l_suppkey": $l.l_suppkey,
+ "l_extendedprice": $l.l_extendedprice,
+ "l_discount": $l.l_discount,
+ "l_quantity": $l.l_quantity,
+ "l_partkey": $l.l_partkey,
+ "l_orderkey": $l.l_orderkey,
+ "n_name": $s1.n_name
+ }
+ )
+ where $ps.ps_suppkey = $l1.l_suppkey and $ps.ps_partkey = $l1.l_partkey
+ return {
+ "l_extendedprice": $l1.l_extendedprice,
+ "l_discount": $l1.l_discount,
+ "l_quantity": $l1.l_quantity,
+ "l_partkey": $l1.l_partkey,
+ "l_orderkey": $l1.l_orderkey,
+ "n_name": $l1.n_name,
+ "ps_supplycost": $ps.ps_supplycost
+ }
+ )
+ where contains($p.p_name, 'green') and $p.p_partkey = $l2.l_partkey
+ return {
+ "l_extendedprice": $l2.l_extendedprice,
+ "l_discount": $l2.l_discount,
+ "l_quantity": $l2.l_quantity,
+ "l_orderkey": $l2.l_orderkey,
+ "n_name": $l2.n_name,
+ "ps_supplycost": $l2.ps_supplycost
+ }
\ No newline at end of file
diff --git a/asterix-app/src/test/resources/optimizerts/results/q08_group_by.plan b/asterix-app/src/test/resources/optimizerts/results/q08_group_by.plan
new file mode 100644
index 0000000..fefdfcb
--- /dev/null
+++ b/asterix-app/src/test/resources/optimizerts/results/q08_group_by.plan
@@ -0,0 +1,78 @@
+-- DISTRIBUTE_RESULT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ASSIGN |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- HYBRID_HASH_JOIN [$$78][$$104] |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ASSIGN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$104] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- HYBRID_HASH_JOIN [$$101][$$85] |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$101] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- HYBRID_HASH_JOIN [$$79][$$81] |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$79] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ASSIGN |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$81] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- HYBRID_HASH_JOIN [$$96][$$82] |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$96] |PARTITIONED|
+ -- STREAM_SELECT |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ASSIGN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$82] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- HYBRID_HASH_JOIN [$$92][$$83] |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$92] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ASSIGN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$83] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- HYBRID_HASH_JOIN [$$89][$$84] |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$89] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ASSIGN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- STREAM_SELECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- STREAM_SELECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|
diff --git a/asterix-app/src/test/resources/optimizerts/results/q09_group_by.plan b/asterix-app/src/test/resources/optimizerts/results/q09_group_by.plan
new file mode 100644
index 0000000..8a08e39
--- /dev/null
+++ b/asterix-app/src/test/resources/optimizerts/results/q09_group_by.plan
@@ -0,0 +1,55 @@
+-- DISTRIBUTE_RESULT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ASSIGN |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- HYBRID_HASH_JOIN [$$62][$$80] |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- STREAM_SELECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$80] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- HYBRID_HASH_JOIN [$$64, $$63][$$69, $$80] |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ASSIGN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$80, $$69] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- HYBRID_HASH_JOIN [$$65][$$69] |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$65] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- HYBRID_HASH_JOIN [$$73][$$66] |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$73] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ASSIGN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ASSIGN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|
+ -- HASH_PARTITION_EXCHANGE [$$69] |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ASSIGN |PARTITIONED|
+ -- STREAM_PROJECT |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- DATASOURCE_SCAN |PARTITIONED|
+ -- ONE_TO_ONE_EXCHANGE |PARTITIONED|
+ -- EMPTY_TUPLE_SOURCE |PARTITIONED|