Test explaining partial pushdown filter with isnull

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

Author: yuancu

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

@@ -5,9 +5,11 @@
 
 package org.opensearch.sql.calcite.remote;
 
+import static org.opensearch.sql.legacy.TestsConstants.TEST_INDEX_NESTED_SIMPLE;
 import static org.opensearch.sql.util.MatcherUtils.assertJsonEqualsIgnoreId;
 
 import java.io.IOException;
+import java.util.Locale;
 import org.junit.Assume;
 import org.junit.Ignore;
 import org.junit.Test;
@@ -18,6 +20,7 @@ public class CalciteExplainIT extends ExplainIT {
   public void init() throws Exception {
     super.init();
     enableCalcite();
+    loadIndex(Index.NESTED_SIMPLE);
   }
 
   @Override
@@ -108,6 +111,20 @@ public void supportPartialPushDownScript() throws IOException {
     assertJsonEqualsIgnoreId(expected, result);
   }
 
+  @Test
+  public void testPartialPushdownFilterWithIsNull() throws IOException {
+    // isnull(nested_field) should not be pushed down since DSL doesn't handle it correctly, but
+    // name='david' can be pushed down
+    String query =
+        String.format(
+            Locale.ROOT,
+            "source=%s | where isnull(address) and name='david'",
+            TEST_INDEX_NESTED_SIMPLE);
+    var result = explainQueryToString(query);
+    String expected = loadExpectedPlan("explain_partial_filter_isnull.json");
+    assertJsonEqualsIgnoreId(expected, result);
+  }
+
   @Test
   public void testSkipScriptEncodingOnExtendedFormat() throws IOException {
     Assume.assumeTrue("This test is only for push down enabled", isPushdownEnabled());

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

@@ -0,0 +1,6 @@
+{
+  "calcite": {
+    "logical": "LogicalSystemLimit(fetch=[10000], type=[QUERY_SIZE_LIMIT])\n  LogicalProject(name=[$0], address=[$1], id=[$6], age=[$7])\n    LogicalFilter(condition=[AND(IS NULL($1), =($0, 'david'))])\n      CalciteLogicalIndexScan(table=[[OpenSearch, opensearch-sql_test_index_nested_simple]])\n",
+    "physical": "EnumerableLimit(fetch=[10000])\n  EnumerableCalc(expr#0..3=[{inputs}], expr#4=[IS NULL($t1)], proj#0..3=[{exprs}], $condition=[$t4])\n    CalciteEnumerableIndexScan(table=[[OpenSearch, opensearch-sql_test_index_nested_simple]], PushDownContext=[[PROJECT->[name, address, id, age], FILTER->=($0, 'david')], OpenSearchRequestBuilder(sourceBuilder={\"from\":0,\"timeout\":\"1m\",\"query\":{\"bool\":{\"must\":[{\"term\":{\"name.keyword\":{\"value\":\"david\",\"boost\":1.0}}}],\"adjust_pure_negative\":true,\"boost\":1.0}},\"_source\":{\"includes\":[\"name\",\"address\",\"id\",\"age\"],\"excludes\":[]},\"sort\":[{\"_doc\":{\"order\":\"asc\"}}]}, requestedTotalSize=2147483647, pageSize=null, startFrom=0)])\n"
+  }
+}

integ-test/src/test/resources/expectedOutput/calcite_no_pushdown/explain_partial_filter_isnull.json

@@ -0,0 +1,6 @@
+{
+  "calcite": {
+    "logical": "LogicalSystemLimit(fetch=[10000], type=[QUERY_SIZE_LIMIT])\n  LogicalProject(name=[$0], address=[$1], id=[$6], age=[$7])\n    LogicalFilter(condition=[AND(IS NULL($1), =($0, 'david'))])\n      CalciteLogicalIndexScan(table=[[OpenSearch, opensearch-sql_test_index_nested_simple]])\n",
+    "physical": "EnumerableLimit(fetch=[10000])\n  EnumerableCalc(expr#0..13=[{inputs}], expr#14=[IS NULL($t1)], expr#15=['david':VARCHAR], expr#16=[=($t0, $t15)], expr#17=[AND($t14, $t16)], proj#0..1=[{exprs}], id=[$t6], age=[$t7], $condition=[$t17])\n    CalciteEnumerableIndexScan(table=[[OpenSearch, opensearch-sql_test_index_nested_simple]])\n"
+  }
+}

opensearch/src/main/java/org/opensearch/sql/opensearch/request/PredicateAnalyzer.java

@@ -1362,6 +1362,8 @@ public ScriptQueryExpression(
         RelDataType rowType,
         Map<String, ExprType> fieldTypes,
         RelOptCluster cluster) {
+      // We prevent is_null(nested_field) from being pushed down because pushed-down scripts can not
+      // access nested fields for the time being
       if (rexNode instanceof RexCall
           && (rexNode.getKind().equals(SqlKind.IS_NULL)
               || rexNode.getKind().equals(SqlKind.IS_NOT_NULL))) {