SQL: Enable accurate hit tracking on demand (#39527)

Queries that require counting of all hits (COUNT(*) on implicit
group by), now enable accurate hit tracking.

Fix #37971

(cherry picked from commit 265b637)

x-pack/plugin/sql/qa/single-node/src/test/java/org/elasticsearch/xpack/sql/qa/single_node/CliExplainIT.java

@@ -153,6 +153,7 @@ public void testExplainWithCount() throws IOException {
         assertThat(readLine(), startsWith("      }"));
         assertThat(readLine(), startsWith("    }"));
         assertThat(readLine(), startsWith("  ]"));
+        assertThat(readLine(), startsWith("  \"track_total_hits\" : 2147483647"));
         assertThat(readLine(), startsWith("}]"));
         assertEquals("", readLine());
     }

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/execution/search/SourceGenerator.java

@@ -169,6 +169,9 @@ private static void optimize(QueryContainer query, SearchSourceBuilder builder)
             // disable source fetching (only doc values are used)
             disableSource(builder);
         }
+        if (query.shouldTrackHits()) {
+            builder.trackTotalHits(true);
+        }
     }
 
     private static void disableSource(SearchSourceBuilder builder) {

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/planner/QueryFolder.java

@@ -14,7 +14,6 @@
 import org.elasticsearch.xpack.sql.expression.Expression;
 import org.elasticsearch.xpack.sql.expression.Expressions;
 import org.elasticsearch.xpack.sql.expression.Foldables;
-import org.elasticsearch.xpack.sql.expression.Literal;
 import org.elasticsearch.xpack.sql.expression.NamedExpression;
 import org.elasticsearch.xpack.sql.expression.Order;
 import org.elasticsearch.xpack.sql.expression.function.Function;
@@ -152,7 +151,8 @@ protected PhysicalPlan rule(ProjectExec project) {
                         queryC.pseudoFunctions(),
                         new AttributeMap<>(processors),
                         queryC.sort(),
-                        queryC.limit());
+                        queryC.limit(),
+                        queryC.shouldTrackHits());
                 return new EsQueryExec(exec.source(), exec.index(), project.output(), clone);
             }
             return project;
@@ -180,7 +180,8 @@ protected PhysicalPlan rule(FilterExec plan) {
                         qContainer.pseudoFunctions(),
                         qContainer.scalarFunctions(),
                         qContainer.sort(),
-                        qContainer.limit());
+                        qContainer.limit(),
+                        qContainer.shouldTrackHits());
 
                 return exec.with(qContainer);
             }
@@ -391,10 +392,16 @@ private Tuple<QueryContainer, AggPathInput> addAggFunction(GroupByKey groupingAg
             if (f instanceof Count) {
                 Count c = (Count) f;
                 // COUNT(*) or COUNT(<literal>)
-                if (c.field() instanceof Literal) {
-                    AggRef ref = groupingAgg == null ?
-                            GlobalCountRef.INSTANCE :
-                            new GroupByRef(groupingAgg.id(), Property.COUNT, null);
+                if (c.field().foldable()) {
+                    AggRef ref = null;
+
+                    if (groupingAgg == null) {
+                        ref = GlobalCountRef.INSTANCE;
+                        // if the count points to the total track hits, enable accurate count retrieval
+                        queryC = queryC.withTrackHits();
+                    } else {
+                        ref = new GroupByRef(groupingAgg.id(), Property.COUNT, null);
+                    }
 
                     Map<String, GroupByKey> pseudoFunctions = new LinkedHashMap<>(queryC.pseudoFunctions());
                     pseudoFunctions.put(functionId, groupingAgg);
@@ -406,7 +413,7 @@ private Tuple<QueryContainer, AggPathInput> addAggFunction(GroupByKey groupingAg
                     queryC = queryC.with(queryC.aggs().addAgg(leafAgg));
                     return new Tuple<>(queryC, a);
                 }
-                // the only variant left - COUNT(DISTINCT) - will be covered by the else branch below
+                // the only variant left - COUNT(DISTINCT) - will be covered by the else branch below as it maps to an aggregation
             }
 
             AggPathInput aggInput = null;

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/querydsl/container/QueryContainer.java

@@ -26,7 +26,6 @@
 import org.elasticsearch.xpack.sql.querydsl.agg.Aggs;
 import org.elasticsearch.xpack.sql.querydsl.agg.GroupByKey;
 import org.elasticsearch.xpack.sql.querydsl.agg.LeafAgg;
-import org.elasticsearch.xpack.sql.querydsl.container.GroupByRef.Property;
 import org.elasticsearch.xpack.sql.querydsl.query.BoolQuery;
 import org.elasticsearch.xpack.sql.querydsl.query.MatchAll;
 import org.elasticsearch.xpack.sql.querydsl.query.NestedQuery;
@@ -81,23 +80,26 @@ public class QueryContainer {
 
     private final Set<Sort> sort;
     private final int limit;
+    private final boolean trackHits;
 
     // computed
     private Boolean aggsOnly;
     private Boolean customSort;
 
     public QueryContainer() {
-        this(null, null, null, null, null, null, null, -1);
+        this(null, null, null, null, null, null, null, -1, false);
     }
 
-    public QueryContainer(Query query,
-            Aggs aggs,
-            List<Tuple<FieldExtraction, ExpressionId>> fields,
+    public QueryContainer(Query query, 
+            Aggs aggs, 
+            List<Tuple<FieldExtraction, 
+            ExpressionId>> fields, 
             AttributeMap<Attribute> aliases,
-            Map<String, GroupByKey> pseudoFunctions,
-            AttributeMap<Pipe> scalarFunctions,
-            Set<Sort> sort,
-            int limit) {
+            Map<String, GroupByKey> pseudoFunctions, 
+            AttributeMap<Pipe> scalarFunctions, 
+            Set<Sort> sort, 
+            int limit, 
+            boolean trackHits) {
         this.query = query;
         this.aggs = aggs == null ? Aggs.EMPTY : aggs;
         this.fields = fields == null || fields.isEmpty() ? emptyList() : fields;
@@ -106,6 +108,7 @@ public QueryContainer(Query query,
         this.scalarFunctions = scalarFunctions == null || scalarFunctions.isEmpty() ? AttributeMap.emptyAttributeMap() : scalarFunctions;
         this.sort = sort == null || sort.isEmpty() ? emptySet() : sort;
         this.limit = limit;
+        this.trackHits = trackHits;
     }
 
     /**
@@ -230,38 +233,46 @@ public boolean hasColumns() {
         return fields.size() > 0;
     }
 
+    public boolean shouldTrackHits() {
+        return trackHits;
+    }
+
     //
     // copy methods
     //
 
     public QueryContainer with(Query q) {
-        return new QueryContainer(q, aggs, fields, aliases, pseudoFunctions, scalarFunctions, sort, limit);
+        return new QueryContainer(q, aggs, fields, aliases, pseudoFunctions, scalarFunctions, sort, limit, trackHits);
     }
 
     public QueryContainer withAliases(AttributeMap<Attribute> a) {
-        return new QueryContainer(query, aggs, fields, a, pseudoFunctions, scalarFunctions, sort, limit);
+        return new QueryContainer(query, aggs, fields, a, pseudoFunctions, scalarFunctions, sort, limit, trackHits);
     }
 
     public QueryContainer withPseudoFunctions(Map<String, GroupByKey> p) {
-        return new QueryContainer(query, aggs, fields, aliases, p, scalarFunctions, sort, limit);
+        return new QueryContainer(query, aggs, fields, aliases, p, scalarFunctions, sort, limit, trackHits);
     }
 
     public QueryContainer with(Aggs a) {
-        return new QueryContainer(query, a, fields, aliases, pseudoFunctions, scalarFunctions, sort, limit);
+        return new QueryContainer(query, a, fields, aliases, pseudoFunctions, scalarFunctions, sort, limit, trackHits);
     }
 
     public QueryContainer withLimit(int l) {
-        return l == limit ? this : new QueryContainer(query, aggs, fields, aliases, pseudoFunctions, scalarFunctions, sort, l);
+        return l == limit ? this : new QueryContainer(query, aggs, fields, aliases, pseudoFunctions, scalarFunctions, sort, l, trackHits);
+    }
+
+    public QueryContainer withTrackHits() {
+        return trackHits ? this : new QueryContainer(query, aggs, fields, aliases, pseudoFunctions, scalarFunctions, sort, limit, true);
     }
 
     public QueryContainer withScalarProcessors(AttributeMap<Pipe> procs) {
-        return new QueryContainer(query, aggs, fields, aliases, pseudoFunctions, procs, sort, limit);
+        return new QueryContainer(query, aggs, fields, aliases, pseudoFunctions, procs, sort, limit, trackHits);
     }
 
     public QueryContainer addSort(Sort sortable) {
         Set<Sort> sort = new LinkedHashSet<>(this.sort);
         sort.add(sortable);
-        return new QueryContainer(query, aggs, fields, aliases, pseudoFunctions, scalarFunctions, sort, limit);
+        return new QueryContainer(query, aggs, fields, aliases, pseudoFunctions, scalarFunctions, sort, limit, trackHits);
     }
 
     private String aliasName(Attribute attr) {
@@ -287,7 +298,7 @@ private Tuple<QueryContainer, FieldExtraction> nestedHitFieldRef(FieldAttribute
                 attr.field().isAggregatable(), attr.parent().name());
         nestedRefs.add(nestedFieldRef);
 
-        return new Tuple<>(new QueryContainer(q, aggs, fields, aliases, pseudoFunctions, scalarFunctions, sort, limit),
+        return new Tuple<>(new QueryContainer(q, aggs, fields, aliases, pseudoFunctions, scalarFunctions, sort, limit, trackHits),
                 nestedFieldRef);
     }
 
@@ -390,7 +401,7 @@ public QueryContainer addColumn(FieldExtraction ref, Attribute attr) {
         ExpressionId id = attr instanceof AggregateFunctionAttribute ? ((AggregateFunctionAttribute) attr).innerId() : attr.id();
         return new QueryContainer(query, aggs, combine(fields, new Tuple<>(ref, id)), aliases, pseudoFunctions,
                 scalarFunctions,
-                sort, limit);
+                sort, limit, trackHits);
     }
 
     public AttributeMap<Pipe> scalarFunctions() {
@@ -401,16 +412,6 @@ public AttributeMap<Pipe> scalarFunctions() {
     // agg methods
     //
 
-    public QueryContainer addAggCount(GroupByKey group, ExpressionId functionId) {
-        FieldExtraction ref = group == null ? GlobalCountRef.INSTANCE : new GroupByRef(group.id(), Property.COUNT, null);
-        Map<String, GroupByKey> pseudoFunctions = new LinkedHashMap<>(this.pseudoFunctions);
-        pseudoFunctions.put(functionId.toString(), group);
-        return new QueryContainer(query, aggs, combine(fields, new Tuple<>(ref, functionId)),
-                aliases,
-                pseudoFunctions,
-                scalarFunctions, sort, limit);
-    }
-
     public QueryContainer addAgg(String groupId, LeafAgg agg) {
         return with(aggs.addAgg(agg));
     }
@@ -465,4 +466,4 @@ public String toString() {
             throw new RuntimeException("error rendering", e);
         }
     }
-}
+}

x-pack/plugin/sql/src/test/java/org/elasticsearch/xpack/sql/execution/search/SourceGeneratorTests.java

@@ -10,6 +10,7 @@
 import org.elasticsearch.search.aggregations.AggregatorFactories.Builder;
 import org.elasticsearch.search.aggregations.bucket.composite.CompositeAggregationBuilder;
 import org.elasticsearch.search.builder.SearchSourceBuilder;
+import org.elasticsearch.search.internal.SearchContext;
 import org.elasticsearch.search.sort.FieldSortBuilder;
 import org.elasticsearch.search.sort.SortOrder;
 import org.elasticsearch.test.ESTestCase;
@@ -111,6 +112,13 @@ public void testNoSort() {
         assertEquals(singletonList(fieldSort("_doc").order(SortOrder.ASC)), sourceBuilder.sorts());
     }
 
+    public void testTrackHits() {
+        SearchSourceBuilder sourceBuilder = SourceGenerator.sourceBuilder(new QueryContainer().withTrackHits(), null,
+                randomIntBetween(1, 10));
+        assertEquals("Should have tracked hits", Integer.valueOf(SearchContext.TRACK_TOTAL_HITS_ACCURATE),
+                sourceBuilder.trackTotalHitsUpTo());
+    }
+
     public void testNoSortIfAgg() {
         QueryContainer container = new QueryContainer()
                 .addGroups(singletonList(new GroupByValue("group_id", "group_column")))

x-pack/plugin/sql/src/test/java/org/elasticsearch/xpack/sql/planner/QueryTranslatorTests.java

@@ -678,4 +678,54 @@ public void testTopHitsAggregationWithTwoArgs() {
                     "{\"date\":{\"order\":\"desc\",\"missing\":\"_last\",\"unmapped_type\":\"date\"}}]}}}}}"));
         }
     }
+
+
+    public void testGlobalCountInImplicitGroupByForcesTrackHits() throws Exception {
+        PhysicalPlan p = optimizeAndPlan("SELECT COUNT(*) FROM test");
+        assertEquals(EsQueryExec.class, p.getClass());
+        EsQueryExec eqe = (EsQueryExec) p;
+        assertTrue("Should be tracking hits", eqe.queryContainer().shouldTrackHits());
+    }
+
+    public void testGlobalCountAllInImplicitGroupByForcesTrackHits() throws Exception {
+        PhysicalPlan p = optimizeAndPlan("SELECT COUNT(ALL *) FROM test");
+        assertEquals(EsQueryExec.class, p.getClass());
+        EsQueryExec eqe = (EsQueryExec) p;
+        assertTrue("Should be tracking hits", eqe.queryContainer().shouldTrackHits());
+    }
+
+    public void testGlobalCountInSpecificGroupByDoesNotForceTrackHits() throws Exception {
+        PhysicalPlan p = optimizeAndPlan("SELECT COUNT(*) FROM test GROUP BY int");
+        assertEquals(EsQueryExec.class, p.getClass());
+        EsQueryExec eqe = (EsQueryExec) p;
+        assertFalse("Should NOT be tracking hits", eqe.queryContainer().shouldTrackHits());
+    }
+
+    public void testFieldAllCountDoesNotTrackHits() throws Exception {
+        PhysicalPlan p = optimizeAndPlan("SELECT COUNT(ALL int) FROM test");
+        assertEquals(EsQueryExec.class, p.getClass());
+        EsQueryExec eqe = (EsQueryExec) p;
+        assertFalse("Should NOT be tracking hits", eqe.queryContainer().shouldTrackHits());
+    }
+
+    public void testFieldCountDoesNotTrackHits() throws Exception {
+        PhysicalPlan p = optimizeAndPlan("SELECT COUNT(int) FROM test");
+        assertEquals(EsQueryExec.class, p.getClass());
+        EsQueryExec eqe = (EsQueryExec) p;
+        assertFalse("Should NOT be tracking hits", eqe.queryContainer().shouldTrackHits());
+    }
+
+    public void testDistinctCountDoesNotTrackHits() throws Exception {
+        PhysicalPlan p = optimizeAndPlan("SELECT COUNT(DISTINCT int) FROM test");
+        assertEquals(EsQueryExec.class, p.getClass());
+        EsQueryExec eqe = (EsQueryExec) p;
+        assertFalse("Should NOT be tracking hits", eqe.queryContainer().shouldTrackHits());
+    }
+
+    public void testNoCountDoesNotTrackHits() throws Exception {
+        PhysicalPlan p = optimizeAndPlan("SELECT int FROM test");
+        assertEquals(EsQueryExec.class, p.getClass());
+        EsQueryExec eqe = (EsQueryExec) p;
+        assertFalse("Should NOT be tracking hits", eqe.queryContainer().shouldTrackHits());
+    }
 }