Fix no pushdown ITs

Signed-off-by: Yuanchun Shen <yuanchu@amazon.com>

Author: yuancu

integ-test/src/test/java/org/opensearch/sql/calcite/remote/CalcitePPLAppendcolIT.java

@@ -26,6 +26,9 @@ public void init() throws Exception {
 
   @Test
   public void testAppendCol() throws IOException {
+    // Although the plans are identical, not pushing down resulting the cnt in the first two rows
+    // being null
+    enabledOnlyWhenPushdownIsEnabled();
     JSONObject actual =
         executeQuery(
             String.format(
@@ -55,6 +58,9 @@ public void testAppendCol() throws IOException {
 
   @Test
   public void testAppendColOverride() throws IOException {
+    // Although the plans are identical, not pushing down resulting the cnt in the first two rows
+    // being null
+    enabledOnlyWhenPushdownIsEnabled();
     JSONObject actual =
         executeQuery(
             String.format(

integ-test/src/test/java/org/opensearch/sql/calcite/remote/CalcitePPLExplainIT.java

@@ -6,6 +6,7 @@
 package org.opensearch.sql.calcite.remote;
 
 import static org.opensearch.sql.util.MatcherUtils.assertJsonEquals;
+import static org.opensearch.sql.util.MatcherUtils.assertYamlEqualsIgnoreId;
 
 import java.io.IOException;
 import org.junit.jupiter.api.Test;
@@ -33,13 +34,9 @@ public void init() throws Exception {
 
   @Test
   public void testExplainCommand() throws IOException {
-    var result = explainQueryToString("source=test | where age = 20 | fields name, age");
-    String expected =
-        !isPushdownDisabled()
-            ? loadFromFile("expectedOutput/calcite/explain_filter_w_pushdown.json")
-            : loadFromFile("expectedOutput/calcite/explain_filter_wo_pushdown.json");
-
-    assertJsonEquals(expected, result);
+    var result = explainQueryYaml("source=test | where age = 20 | fields name, age");
+    String expected = loadExpectedPlan("explain_filter.yaml");
+    assertYamlEqualsIgnoreId(expected, result);
   }
 
   @Test

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

@@ -0,0 +1,8 @@
+calcite:
+  logical: |
+    LogicalSystemLimit(fetch=[10000], type=[QUERY_SIZE_LIMIT])
+      LogicalProject(name=[$0], age=[$1])
+        LogicalFilter(condition=[=($1, SAFE_CAST(20:BIGINT))])
+          CalciteLogicalIndexScan(table=[[OpenSearch, test]])
+  physical: |
+    CalciteEnumerableIndexScan(table=[[OpenSearch, test]], PushDownContext=[[PROJECT->[name, age], FILTER->=($1, 20), LIMIT->10000], OpenSearchRequestBuilder(sourceBuilder={"from":0,"size":10000,"timeout":"1m","query":{"term":{"age":{"value":20,"boost":1.0}}},"_source":{"includes":["name","age"],"excludes":[]}}, requestedTotalSize=10000, pageSize=null, startFrom=0)])

integ-test/src/test/resources/expectedOutput/calcite/explain_filter_w_pushdown.json

@@ -3,11 +3,11 @@ calcite:
     LogicalSystemLimit(fetch=[10000], type=[QUERY_SIZE_LIMIT])
       LogicalProject(avg_age=[$1], age_range=[$0])
         LogicalAggregate(group=[{0}], avg_age=[AVG($1)])
-          LogicalProject($f0=[CASE(<($10, 30), 'u30':VARCHAR, SEARCH($10, Sarg[[30..40]]), 'u40':VARCHAR, 'u100':VARCHAR)], age=[$10])
+          LogicalProject(age_range=[CASE(<($10, 30), 'u30':VARCHAR, SEARCH($10, Sarg[[30..40]]), 'u40':VARCHAR, 'u100':VARCHAR)], age=[$10])
             CalciteLogicalIndexScan(table=[[OpenSearch, opensearch-sql_test_index_bank]])
   physical: |
     EnumerableLimit(fetch=[10000])
       EnumerableCalc(expr#0..2=[{inputs}], expr#3=[0], expr#4=[=($t2, $t3)], expr#5=[null:BIGINT], expr#6=[CASE($t4, $t5, $t1)], expr#7=[CAST($t6):DOUBLE], expr#8=[/($t7, $t2)], avg_age=[$t8], age_range=[$t0])
         EnumerableAggregate(group=[{0}], agg#0=[$SUM0($1)], agg#1=[COUNT($1)])
-          EnumerableCalc(expr#0..18=[{inputs}], expr#19=[30], expr#20=[<($t10, $t19)], expr#21=['u30':VARCHAR], expr#22=[Sarg[[30..40]]], expr#23=[SEARCH($t10, $t22)], expr#24=['u40':VARCHAR], expr#25=['u100':VARCHAR], expr#26=[CASE($t20, $t21, $t23, $t24, $t25)], $f0=[$t26], age=[$t10])
+          EnumerableCalc(expr#0..18=[{inputs}], expr#19=[30], expr#20=[<($t10, $t19)], expr#21=['u30':VARCHAR], expr#22=[Sarg[[30..40]]], expr#23=[SEARCH($t10, $t22)], expr#24=['u40':VARCHAR], expr#25=['u100':VARCHAR], expr#26=[CASE($t20, $t21, $t23, $t24, $t25)], age_range=[$t26], age=[$t10])
             CalciteEnumerableIndexScan(table=[[OpenSearch, opensearch-sql_test_index_bank]])

integ-test/src/test/resources/expectedOutput/calcite/explain_filter_wo_pushdown.json

@@ -3,11 +3,11 @@ calcite:
     LogicalSystemLimit(fetch=[10000], type=[QUERY_SIZE_LIMIT])
       LogicalProject(avg_balance=[$2], age_range=[$0], state=[$1])
         LogicalAggregate(group=[{0, 1}], avg_balance=[AVG($2)])
-          LogicalProject($f0=[CASE(<($10, 35), 'u35':VARCHAR, $11)], state=[$9], balance=[$7])
+          LogicalProject(age_range=[CASE(<($10, 35), 'u35':VARCHAR, $11)], state=[$9], balance=[$7])
             CalciteLogicalIndexScan(table=[[OpenSearch, opensearch-sql_test_index_bank]])
   physical: |
     EnumerableLimit(fetch=[10000])
       EnumerableCalc(expr#0..3=[{inputs}], expr#4=[0], expr#5=[=($t3, $t4)], expr#6=[null:BIGINT], expr#7=[CASE($t5, $t6, $t2)], expr#8=[CAST($t7):DOUBLE], expr#9=[/($t8, $t3)], avg_balance=[$t9], age_range=[$t0], state=[$t1])
         EnumerableAggregate(group=[{0, 1}], agg#0=[$SUM0($2)], agg#1=[COUNT($2)])
-          EnumerableCalc(expr#0..18=[{inputs}], expr#19=[35], expr#20=[<($t10, $t19)], expr#21=['u35':VARCHAR], expr#22=[CASE($t20, $t21, $t11)], $f0=[$t22], state=[$t9], balance=[$t7])
+          EnumerableCalc(expr#0..18=[{inputs}], expr#19=[35], expr#20=[<($t10, $t19)], expr#21=['u35':VARCHAR], expr#22=[CASE($t20, $t21, $t11)], age_range=[$t22], state=[$t9], balance=[$t7])
             CalciteEnumerableIndexScan(table=[[OpenSearch, opensearch-sql_test_index_bank]])

integ-test/src/test/resources/expectedOutput/calcite_no_pushdown/agg_case_cannot_push.yaml

@@ -3,11 +3,11 @@ calcite:
     LogicalSystemLimit(fetch=[10000], type=[QUERY_SIZE_LIMIT])
       LogicalProject(count()=[$1], age_range=[$0])
         LogicalAggregate(group=[{0}], count()=[COUNT()])
-          LogicalProject($f0=[CASE(<($10, 30), 30, 100)])
+          LogicalProject(age_range=[CASE(<($10, 30), 30, 100)])
             CalciteLogicalIndexScan(table=[[OpenSearch, opensearch-sql_test_index_bank]])
   physical: |
     EnumerableLimit(fetch=[10000])
       EnumerableCalc(expr#0..1=[{inputs}], count()=[$t1], age_range=[$t0])
         EnumerableAggregate(group=[{0}], count()=[COUNT()])
-          EnumerableCalc(expr#0..18=[{inputs}], expr#19=[30], expr#20=[<($t10, $t19)], expr#21=[100], expr#22=[CASE($t20, $t19, $t21)], $f0=[$t22])
+          EnumerableCalc(expr#0..18=[{inputs}], expr#19=[30], expr#20=[<($t10, $t19)], expr#21=[100], expr#22=[CASE($t20, $t19, $t21)], age_range=[$t22])
             CalciteEnumerableIndexScan(table=[[OpenSearch, opensearch-sql_test_index_bank]])

integ-test/src/test/resources/expectedOutput/calcite_no_pushdown/agg_case_composite_cannot_push.yaml

@@ -3,11 +3,11 @@ calcite:
     LogicalSystemLimit(fetch=[10000], type=[QUERY_SIZE_LIMIT])
       LogicalProject(avg(balance)=[$3], count()=[$4], age_range=[$0], state=[$1], gender=[$2])
         LogicalAggregate(group=[{0, 1, 2}], avg(balance)=[AVG($3)], count()=[COUNT()])
-          LogicalProject($f0=[CASE(<($10, 30), 'u30':VARCHAR, 'a30':VARCHAR)], state=[$9], gender=[$4], balance=[$7])
+          LogicalProject(age_range=[CASE(<($10, 30), 'u30':VARCHAR, 'a30':VARCHAR)], state=[$9], gender=[$4], balance=[$7])
             CalciteLogicalIndexScan(table=[[OpenSearch, opensearch-sql_test_index_bank]])
   physical: |
     EnumerableLimit(fetch=[10000])
       EnumerableCalc(expr#0..5=[{inputs}], expr#6=[0], expr#7=[=($t4, $t6)], expr#8=[null:BIGINT], expr#9=[CASE($t7, $t8, $t3)], expr#10=[CAST($t9):DOUBLE], expr#11=[/($t10, $t4)], avg(balance)=[$t11], count()=[$t5], age_range=[$t0], state=[$t1], gender=[$t2])
         EnumerableAggregate(group=[{0, 1, 2}], agg#0=[$SUM0($3)], agg#1=[COUNT($3)], count()=[COUNT()])
-          EnumerableCalc(expr#0..18=[{inputs}], expr#19=[30], expr#20=[<($t10, $t19)], expr#21=['u30':VARCHAR], expr#22=['a30':VARCHAR], expr#23=[CASE($t20, $t21, $t22)], $f0=[$t23], state=[$t9], gender=[$t4], balance=[$t7])
+          EnumerableCalc(expr#0..18=[{inputs}], expr#19=[30], expr#20=[<($t10, $t19)], expr#21=['u30':VARCHAR], expr#22=['a30':VARCHAR], expr#23=[CASE($t20, $t21, $t22)], age_range=[$t23], state=[$t9], gender=[$t4], balance=[$t7])
             CalciteEnumerableIndexScan(table=[[OpenSearch, opensearch-sql_test_index_bank]])

integ-test/src/test/resources/expectedOutput/calcite_no_pushdown/agg_case_num_res_cannot_push.yaml

@@ -3,13 +3,13 @@ calcite:
     LogicalSystemLimit(fetch=[10000], type=[QUERY_SIZE_LIMIT])
       LogicalProject(avg(value)=[$2], span(@timestamp,1h)=[$1], value_range=[$0])
         LogicalAggregate(group=[{0, 1}], avg(value)=[AVG($2)])
-          LogicalProject(value_range=[$10], $f1=[SPAN($0, 1, 'h')], value=[$2])
+          LogicalProject(value_range=[$10], span(@timestamp,1h)=[SPAN($0, 1, 'h')], value=[$2])
             LogicalFilter(condition=[IS NOT NULL($0)])
               LogicalProject(@timestamp=[$0], category=[$1], value=[$2], timestamp=[$3], _id=[$4], _index=[$5], _score=[$6], _maxscore=[$7], _sort=[$8], _routing=[$9], value_range=[CASE(<($2, 7000), 'small':VARCHAR, 'large':VARCHAR)])
                 CalciteLogicalIndexScan(table=[[OpenSearch, opensearch-sql_test_index_time_data]])
   physical: |
     EnumerableLimit(fetch=[10000])
       EnumerableCalc(expr#0..3=[{inputs}], expr#4=[0], expr#5=[=($t3, $t4)], expr#6=[null:BIGINT], expr#7=[CASE($t5, $t6, $t2)], expr#8=[CAST($t7):DOUBLE], expr#9=[/($t8, $t3)], avg(value)=[$t9], span(@timestamp,1h)=[$t1], value_range=[$t0])
         EnumerableAggregate(group=[{0, 1}], agg#0=[$SUM0($2)], agg#1=[COUNT($2)])
-          EnumerableCalc(expr#0..9=[{inputs}], expr#10=[7000], expr#11=[<($t2, $t10)], expr#12=['small':VARCHAR], expr#13=['large':VARCHAR], expr#14=[CASE($t11, $t12, $t13)], expr#15=[1], expr#16=['h'], expr#17=[SPAN($t0, $t15, $t16)], expr#18=[IS NOT NULL($t0)], value_range=[$t14], $f1=[$t17], value=[$t2], $condition=[$t18])
+          EnumerableCalc(expr#0..9=[{inputs}], expr#10=[7000], expr#11=[<($t2, $t10)], expr#12=['small':VARCHAR], expr#13=['large':VARCHAR], expr#14=[CASE($t11, $t12, $t13)], expr#15=[1], expr#16=['h'], expr#17=[SPAN($t0, $t15, $t16)], expr#18=[IS NOT NULL($t0)], value_range=[$t14], span(@timestamp,1h)=[$t17], value=[$t2], $condition=[$t18])
             CalciteEnumerableIndexScan(table=[[OpenSearch, opensearch-sql_test_index_time_data]])