Update expected plans for timechart

Signed-off-by: Yuanchun Shen <[email protected]>

Author: yuancu

integ-test/src/test/resources/expectedOutput/calcite/explain_timechart.yaml

@@ -3,37 +3,38 @@ calcite:
     LogicalSystemLimit(sort0=[$0], sort1=[$1], dir0=[ASC], dir1=[ASC], fetch=[10000], type=[QUERY_SIZE_LIMIT])
       LogicalSort(sort0=[$0], sort1=[$1], dir0=[ASC], dir1=[ASC])
         LogicalAggregate(group=[{0, 1}], avg(cpu_usage)=[AVG($2)])
-          LogicalProject(@timestamp=[$0], host=[CASE(IS NOT NULL($3), $1, CASE(IS NULL($1), null:NULL, 'OTHER'))], avg(cpu_usage)=[$2])
+          LogicalProject(@timestamp=[$0], host=[CASE(IS NULL($1), 'NULL', <=($5, 10), $1, 'OTHER')], avg(cpu_usage)=[$2])
             LogicalJoin(condition=[=($1, $3)], joinType=[left])
-              LogicalProject(@timestamp=[$1], host=[$0], $f2=[$2])
-                LogicalAggregate(group=[{0, 2}], agg#0=[AVG($1)])
-                  LogicalProject(host=[$4], cpu_usage=[$7], $f3=[SPAN($1, 1, 'm')])
-                    CalciteLogicalIndexScan(table=[[OpenSearch, events]])
-              LogicalSort(sort0=[$1], dir0=[DESC], fetch=[10])
-                LogicalAggregate(group=[{1}], grand_total=[AVG($2)])
-                  LogicalFilter(condition=[IS NOT NULL($1)])
-                    LogicalProject(@timestamp=[$1], host=[$0], $f2=[$2])
-                      LogicalAggregate(group=[{0, 2}], agg#0=[AVG($1)])
-                        LogicalProject(host=[$4], cpu_usage=[$7], $f3=[SPAN($1, 1, 'm')])
-                          CalciteLogicalIndexScan(table=[[OpenSearch, events]])
+              LogicalProject(@timestamp=[$1], host=[$0], avg(cpu_usage)=[$2])
+                LogicalAggregate(group=[{0, 2}], avg(cpu_usage)=[AVG($1)])
+                  LogicalProject(host=[$4], cpu_usage=[$7], @timestamp0=[SPAN($1, 1, 'm')])
+                    LogicalFilter(condition=[AND(IS NOT NULL($1), IS NOT NULL($7))])
+                      CalciteLogicalIndexScan(table=[[OpenSearch, events]])
+              LogicalProject(host=[$0], __grand_total__=[$1], _row_number_chart_=[ROW_NUMBER() OVER (ORDER BY $1 DESC NULLS LAST)])
+                LogicalAggregate(group=[{0}], __grand_total__=[SUM($1)])
+                  LogicalFilter(condition=[IS NOT NULL($0)])
+                    LogicalProject(host=[$0], avg(cpu_usage)=[$2])
+                      LogicalAggregate(group=[{0, 2}], avg(cpu_usage)=[AVG($1)])
+                        LogicalProject(host=[$4], cpu_usage=[$7], @timestamp0=[SPAN($1, 1, 'm')])
+                          LogicalFilter(condition=[AND(IS NOT NULL($1), IS NOT NULL($7))])
+                            CalciteLogicalIndexScan(table=[[OpenSearch, events]])
   physical: |
     EnumerableLimit(fetch=[10000])
       EnumerableSort(sort0=[$0], sort1=[$1], dir0=[ASC], dir1=[ASC])
         EnumerableCalc(expr#0..3=[{inputs}], expr#4=[0], expr#5=[=($t3, $t4)], expr#6=[null:DOUBLE], expr#7=[CASE($t5, $t6, $t2)], expr#8=[/($t7, $t3)], proj#0..1=[{exprs}], avg(cpu_usage)=[$t8])
           EnumerableAggregate(group=[{0, 1}], agg#0=[$SUM0($2)], agg#1=[COUNT($2)])
-            EnumerableCalc(expr#0..4=[{inputs}], expr#5=[IS NOT NULL($t3)], expr#6=[IS NULL($t1)], expr#7=[null:NULL], expr#8=['OTHER'], expr#9=[CASE($t6, $t7, $t8)], expr#10=[CASE($t5, $t1, $t9)], @timestamp=[$t0], host=[$t10], avg(cpu_usage)=[$t2])
+            EnumerableCalc(expr#0..4=[{inputs}], expr#5=[IS NULL($t1)], expr#6=['NULL'], expr#7=[10], expr#8=[<=($t4, $t7)], expr#9=['OTHER'], expr#10=[CASE($t5, $t6, $t8, $t1, $t9)], @timestamp=[$t0], host=[$t10], avg(cpu_usage)=[$t2])
               EnumerableMergeJoin(condition=[=($1, $3)], joinType=[left])
                 EnumerableSort(sort0=[$1], dir0=[ASC])
-                  EnumerableCalc(expr#0..3=[{inputs}], expr#4=[0], expr#5=[=($t3, $t4)], expr#6=[null:DOUBLE], expr#7=[CASE($t5, $t6, $t2)], expr#8=[/($t7, $t3)], @timestamp=[$t1], host=[$t0], $f2=[$t8])
+                  EnumerableCalc(expr#0..3=[{inputs}], expr#4=[0], expr#5=[=($t3, $t4)], expr#6=[null:DOUBLE], expr#7=[CASE($t5, $t6, $t2)], expr#8=[/($t7, $t3)], @timestamp=[$t1], host=[$t0], avg(cpu_usage)=[$t8])
                     EnumerableAggregate(group=[{0, 2}], agg#0=[$SUM0($1)], agg#1=[COUNT($1)])
-                      EnumerableCalc(expr#0..2=[{inputs}], expr#3=[1], expr#4=['m'], expr#5=[SPAN($t2, $t3, $t4)], proj#0..1=[{exprs}], $f2=[$t5])
-                        CalciteEnumerableIndexScan(table=[[OpenSearch, events]], PushDownContext=[[PROJECT->[host, cpu_usage, @timestamp]], OpenSearchRequestBuilder(sourceBuilder={"from":0,"timeout":"1m","_source":{"includes":["host","cpu_usage","@timestamp"],"excludes":[]}}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])
+                      EnumerableCalc(expr#0..2=[{inputs}], expr#3=[1], expr#4=['m'], expr#5=[SPAN($t0, $t3, $t4)], host=[$t1], cpu_usage=[$t2], @timestamp0=[$t5])
+                        CalciteEnumerableIndexScan(table=[[OpenSearch, events]], PushDownContext=[[PROJECT->[@timestamp, host, cpu_usage], FILTER->AND(IS NOT NULL($0), IS NOT NULL($2))], OpenSearchRequestBuilder(sourceBuilder={"from":0,"timeout":"1m","query":{"bool":{"must":[{"exists":{"field":"@timestamp","boost":1.0}},{"exists":{"field":"cpu_usage","boost":1.0}}],"adjust_pure_negative":true,"boost":1.0}},"_source":{"includes":["@timestamp","host","cpu_usage"],"excludes":[]}}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])
                 EnumerableSort(sort0=[$0], dir0=[ASC])
-                  EnumerableLimit(fetch=[10])
-                    EnumerableSort(sort0=[$1], dir0=[DESC])
-                      EnumerableCalc(expr#0..2=[{inputs}], expr#3=[0], expr#4=[=($t2, $t3)], expr#5=[null:DOUBLE], expr#6=[CASE($t4, $t5, $t1)], expr#7=[/($t6, $t2)], host=[$t0], grand_total=[$t7])
-                        EnumerableAggregate(group=[{0}], agg#0=[$SUM0($1)], agg#1=[COUNT($1)])
-                          EnumerableCalc(expr#0..3=[{inputs}], expr#4=[0], expr#5=[=($t3, $t4)], expr#6=[null:DOUBLE], expr#7=[CASE($t5, $t6, $t2)], expr#8=[/($t7, $t3)], host=[$t0], $f2=[$t8])
-                            EnumerableAggregate(group=[{0, 2}], agg#0=[$SUM0($1)], agg#1=[COUNT($1)])
-                              EnumerableCalc(expr#0..2=[{inputs}], expr#3=[1], expr#4=['m'], expr#5=[SPAN($t2, $t3, $t4)], proj#0..1=[{exprs}], $f2=[$t5])
-                                CalciteEnumerableIndexScan(table=[[OpenSearch, events]], PushDownContext=[[PROJECT->[host, cpu_usage, @timestamp], FILTER->IS NOT NULL($0)], OpenSearchRequestBuilder(sourceBuilder={"from":0,"timeout":"1m","query":{"exists":{"field":"host","boost":1.0}},"_source":{"includes":["host","cpu_usage","@timestamp"],"excludes":[]}}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])
+                  EnumerableCalc(expr#0..2=[{inputs}], host=[$t0], $1=[$t2])
+                    EnumerableWindow(window#0=[window(order by [1 DESC-nulls-last] rows between UNBOUNDED PRECEDING and CURRENT ROW aggs [ROW_NUMBER()])])
+                      EnumerableAggregate(group=[{0}], __grand_total__=[SUM($1)])
+                        EnumerableCalc(expr#0..3=[{inputs}], expr#4=[0], expr#5=[=($t3, $t4)], expr#6=[null:DOUBLE], expr#7=[CASE($t5, $t6, $t2)], expr#8=[/($t7, $t3)], host=[$t0], avg(cpu_usage)=[$t8])
+                          EnumerableAggregate(group=[{0, 2}], agg#0=[$SUM0($1)], agg#1=[COUNT($1)])
+                            EnumerableCalc(expr#0..2=[{inputs}], expr#3=[1], expr#4=['m'], expr#5=[SPAN($t0, $t3, $t4)], host=[$t1], cpu_usage=[$t2], @timestamp0=[$t5])
+                              CalciteEnumerableIndexScan(table=[[OpenSearch, events]], PushDownContext=[[PROJECT->[@timestamp, host, cpu_usage], FILTER->AND(IS NOT NULL($0), IS NOT NULL($2)), FILTER->IS NOT NULL($1)], OpenSearchRequestBuilder(sourceBuilder={"from":0,"timeout":"1m","query":{"bool":{"filter":[{"bool":{"must":[{"exists":{"field":"@timestamp","boost":1.0}},{"exists":{"field":"cpu_usage","boost":1.0}}],"adjust_pure_negative":true,"boost":1.0}},{"exists":{"field":"host","boost":1.0}}],"adjust_pure_negative":true,"boost":1.0}},"_source":{"includes":["@timestamp","host","cpu_usage"],"excludes":[]}}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])

integ-test/src/test/resources/expectedOutput/calcite/explain_timechart_count.yaml

@@ -2,70 +2,35 @@ calcite:
   logical: |
     LogicalSystemLimit(sort0=[$0], sort1=[$1], dir0=[ASC], dir1=[ASC], fetch=[10000], type=[QUERY_SIZE_LIMIT])
       LogicalSort(sort0=[$0], sort1=[$1], dir0=[ASC], dir1=[ASC])
-        LogicalAggregate(group=[{0, 1}], count=[SUM($2)])
-          LogicalUnion(all=[false])
-            LogicalAggregate(group=[{0, 1}], actual_count=[SUM($2)])
-              LogicalProject(@timestamp=[CAST($0):TIMESTAMP(0) NOT NULL], host=[CASE(IS NOT NULL($3), $1, CASE(IS NULL($1), null:NULL, 'OTHER'))], count=[$2])
-                LogicalJoin(condition=[IS NOT DISTINCT FROM($1, $3)], joinType=[left])
-                  LogicalProject(@timestamp=[$1], host=[$0], $f2_0=[$2])
-                    LogicalAggregate(group=[{0, 1}], agg#0=[COUNT()])
-                      LogicalProject(host=[$4], $f2=[SPAN($1, 1, 'm')])
-                        CalciteLogicalIndexScan(table=[[OpenSearch, events]])
-                  LogicalSort(sort0=[$1], dir0=[DESC], fetch=[10])
-                    LogicalAggregate(group=[{1}], grand_total=[SUM($2)])
-                      LogicalFilter(condition=[IS NOT NULL($1)])
-                        LogicalProject(@timestamp=[$1], host=[$0], $f2_0=[$2])
-                          LogicalAggregate(group=[{0, 1}], agg#0=[COUNT()])
-                            LogicalProject(host=[$4], $f2=[SPAN($1, 1, 'm')])
-                              CalciteLogicalIndexScan(table=[[OpenSearch, events]])
-            LogicalProject(@timestamp=[CAST($0):TIMESTAMP(0) NOT NULL], host=[$1], count=[0])
-              LogicalJoin(condition=[true], joinType=[inner])
-                LogicalAggregate(group=[{0}])
-                  LogicalProject(@timestamp=[$1])
-                    LogicalAggregate(group=[{0, 1}], agg#0=[COUNT()])
-                      LogicalProject(host=[$4], $f2=[SPAN($1, 1, 'm')])
-                        CalciteLogicalIndexScan(table=[[OpenSearch, events]])
-                LogicalAggregate(group=[{0}])
-                  LogicalProject($f0=[CASE(IS NOT NULL($3), $1, CASE(IS NULL($1), null:NULL, 'OTHER'))])
-                    LogicalJoin(condition=[IS NOT DISTINCT FROM($1, $3)], joinType=[left])
-                      LogicalProject(@timestamp=[$1], host=[$0], $f2_0=[$2])
-                        LogicalAggregate(group=[{0, 1}], agg#0=[COUNT()])
-                          LogicalProject(host=[$4], $f2=[SPAN($1, 1, 'm')])
-                            CalciteLogicalIndexScan(table=[[OpenSearch, events]])
-                      LogicalSort(sort0=[$1], dir0=[DESC], fetch=[10])
-                        LogicalAggregate(group=[{1}], grand_total=[SUM($2)])
+        LogicalAggregate(group=[{0, 1}], count()=[SUM($2)])
+          LogicalProject(@timestamp=[$0], host=[CASE(IS NULL($1), 'NULL', <=($5, 10), $1, 'OTHER')], count()=[$2])
+            LogicalJoin(condition=[=($1, $3)], joinType=[left])
+              LogicalProject(@timestamp=[$1], host=[$0], count()=[$2])
+                LogicalAggregate(group=[{0, 1}], count()=[COUNT()])
+                  LogicalProject(host=[$4], @timestamp0=[SPAN($1, 1, 'm')])
+                    LogicalFilter(condition=[IS NOT NULL($1)])
+                      CalciteLogicalIndexScan(table=[[OpenSearch, events]])
+              LogicalProject(host=[$0], __grand_total__=[$1], _row_number_chart_=[ROW_NUMBER() OVER (ORDER BY $1 DESC NULLS LAST)])
+                LogicalAggregate(group=[{0}], __grand_total__=[SUM($1)])
+                  LogicalFilter(condition=[IS NOT NULL($0)])
+                    LogicalProject(host=[$0], count()=[$2])
+                      LogicalAggregate(group=[{0, 1}], count()=[COUNT()])
+                        LogicalProject(host=[$4], @timestamp0=[SPAN($1, 1, 'm')])
                           LogicalFilter(condition=[IS NOT NULL($1)])
-                            LogicalProject(@timestamp=[$1], host=[$0], $f2_0=[$2])
-                              LogicalAggregate(group=[{0, 1}], agg#0=[COUNT()])
-                                LogicalProject(host=[$4], $f2=[SPAN($1, 1, 'm')])
-                                  CalciteLogicalIndexScan(table=[[OpenSearch, events]])
+                            CalciteLogicalIndexScan(table=[[OpenSearch, events]])
   physical: |
     EnumerableLimit(fetch=[10000])
       EnumerableSort(sort0=[$0], sort1=[$1], dir0=[ASC], dir1=[ASC])
-        EnumerableAggregate(group=[{0, 1}], count=[$SUM0($2)])
-          EnumerableUnion(all=[false])
-            EnumerableAggregate(group=[{0, 1}], actual_count=[$SUM0($2)])
-              EnumerableCalc(expr#0..4=[{inputs}], expr#5=[CAST($t0):TIMESTAMP(0) NOT NULL], expr#6=[IS NOT NULL($t3)], expr#7=[IS NULL($t1)], expr#8=[null:NULL], expr#9=['OTHER'], expr#10=[CASE($t7, $t8, $t9)], expr#11=[CASE($t6, $t1, $t10)], @timestamp=[$t5], host=[$t11], count=[$t2])
-                EnumerableNestedLoopJoin(condition=[IS NOT DISTINCT FROM($1, $3)], joinType=[left])
-                  EnumerableCalc(expr#0..2=[{inputs}], @timestamp=[$t1], host=[$t0], $f2_0=[$t2])
-                    EnumerableAggregate(group=[{0, 1}], agg#0=[COUNT()])
-                      EnumerableCalc(expr#0..1=[{inputs}], expr#2=[1], expr#3=['m'], expr#4=[SPAN($t1, $t2, $t3)], host=[$t0], $f1=[$t4])
-                        CalciteEnumerableIndexScan(table=[[OpenSearch, events]], PushDownContext=[[PROJECT->[host, @timestamp]], OpenSearchRequestBuilder(sourceBuilder={"from":0,"timeout":"1m","_source":{"includes":["host","@timestamp"],"excludes":[]}}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])
-                  EnumerableLimit(fetch=[10])
-                    EnumerableSort(sort0=[$1], dir0=[DESC])
-                      EnumerableAggregate(group=[{0}], grand_total=[COUNT()])
-                        CalciteEnumerableIndexScan(table=[[OpenSearch, events]], PushDownContext=[[PROJECT->[host, @timestamp], FILTER->IS NOT NULL($0)], OpenSearchRequestBuilder(sourceBuilder={"from":0,"timeout":"1m","query":{"exists":{"field":"host","boost":1.0}},"_source":{"includes":["host","@timestamp"],"excludes":[]}}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])
-            EnumerableCalc(expr#0..1=[{inputs}], expr#2=[CAST($t0):TIMESTAMP(0) NOT NULL], expr#3=[0], @timestamp=[$t2], host=[$t1], count=[$t3])
-              EnumerableNestedLoopJoin(condition=[true], joinType=[inner])
-                CalciteEnumerableIndexScan(table=[[OpenSearch, events]], PushDownContext=[[AGGREGATION->rel#:LogicalAggregate.NONE.[](input=RelSubset#,group={1})], OpenSearchRequestBuilder(sourceBuilder={"from":0,"size":0,"timeout":"1m","aggregations":{"composite_buckets":{"composite":{"size":1000,"sources":[{"$f2":{"date_histogram":{"field":"@timestamp","missing_bucket":false,"order":"asc","fixed_interval":"1m"}}}]}}}}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])
-                EnumerableAggregate(group=[{0}])
-                  EnumerableCalc(expr#0..2=[{inputs}], expr#3=[IS NOT NULL($t1)], expr#4=[IS NULL($t0)], expr#5=[null:NULL], expr#6=['OTHER'], expr#7=[CASE($t4, $t5, $t6)], expr#8=[CASE($t3, $t0, $t7)], $f0=[$t8])
-                    EnumerableNestedLoopJoin(condition=[IS NOT DISTINCT FROM($0, $1)], joinType=[left])
-                      EnumerableCalc(expr#0..1=[{inputs}], host=[$t0])
-                        EnumerableAggregate(group=[{0, 1}])
-                          EnumerableCalc(expr#0..1=[{inputs}], expr#2=[1], expr#3=['m'], expr#4=[SPAN($t1, $t2, $t3)], host=[$t0], $f1=[$t4])
-                            CalciteEnumerableIndexScan(table=[[OpenSearch, events]], PushDownContext=[[PROJECT->[host, @timestamp]], OpenSearchRequestBuilder(sourceBuilder={"from":0,"timeout":"1m","_source":{"includes":["host","@timestamp"],"excludes":[]}}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])
-                      EnumerableLimit(fetch=[10])
-                        EnumerableSort(sort0=[$1], dir0=[DESC])
-                          EnumerableAggregate(group=[{0}], grand_total=[COUNT()])
-                            CalciteEnumerableIndexScan(table=[[OpenSearch, events]], PushDownContext=[[PROJECT->[host, @timestamp], FILTER->IS NOT NULL($0)], OpenSearchRequestBuilder(sourceBuilder={"from":0,"timeout":"1m","query":{"exists":{"field":"host","boost":1.0}},"_source":{"includes":["host","@timestamp"],"excludes":[]}}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])
+        EnumerableAggregate(group=[{0, 1}], count()=[$SUM0($2)])
+          EnumerableCalc(expr#0..4=[{inputs}], expr#5=[IS NULL($t1)], expr#6=['NULL'], expr#7=[10], expr#8=[<=($t4, $t7)], expr#9=['OTHER'], expr#10=[CASE($t5, $t6, $t8, $t1, $t9)], @timestamp=[$t0], host=[$t10], count()=[$t2])
+            EnumerableMergeJoin(condition=[=($1, $3)], joinType=[left])
+              EnumerableSort(sort0=[$1], dir0=[ASC])
+                EnumerableCalc(expr#0..2=[{inputs}], @timestamp=[$t1], host=[$t0], count()=[$t2])
+                  EnumerableAggregate(group=[{0, 1}], count()=[COUNT()])
+                    EnumerableCalc(expr#0..1=[{inputs}], expr#2=[1], expr#3=['m'], expr#4=[SPAN($t0, $t2, $t3)], host=[$t1], @timestamp0=[$t4])
+                      CalciteEnumerableIndexScan(table=[[OpenSearch, events]], PushDownContext=[[PROJECT->[@timestamp, host], FILTER->IS NOT NULL($0)], OpenSearchRequestBuilder(sourceBuilder={"from":0,"timeout":"1m","query":{"exists":{"field":"@timestamp","boost":1.0}},"_source":{"includes":["@timestamp","host"],"excludes":[]}}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])
+              EnumerableSort(sort0=[$0], dir0=[ASC])
+                EnumerableCalc(expr#0..2=[{inputs}], host=[$t0], $1=[$t2])
+                  EnumerableWindow(window#0=[window(order by [1 DESC-nulls-last] rows between UNBOUNDED PRECEDING and CURRENT ROW aggs [ROW_NUMBER()])])
+                    EnumerableAggregate(group=[{1}], __grand_total__=[COUNT()])
+                      CalciteEnumerableIndexScan(table=[[OpenSearch, events]], PushDownContext=[[PROJECT->[@timestamp, host], FILTER->IS NOT NULL($0), FILTER->IS NOT NULL($1)], OpenSearchRequestBuilder(sourceBuilder={"from":0,"timeout":"1m","query":{"bool":{"filter":[{"exists":{"field":"@timestamp","boost":1.0}},{"exists":{"field":"host","boost":1.0}}],"adjust_pure_negative":true,"boost":1.0}},"_source":{"includes":["@timestamp","host"],"excludes":[]}}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])