SQL: Failing Group By queries due to different ExpressionIds (#43072)

Fix an issue that arises from the use of ExpressionIds as keys in a lookup map
that helps the QueryTranslator to identify the grouping columns. The issue is
that the same expression in different parts of the query (SELECT clause and GROUP BY clause)
ends up with different ExpressionIds so the lookup fails. So, instead of ExpressionIds
use the hashCode() of NamedExpression.

Fixes: #41159
Fixes: #40001
Fixes: #40240
Fixes: #33361
Fixes: #46316
Fixes: #36074
Fixes: #34543
Fixes: #37044

Fixes: #42041
(cherry picked from commit 3c38ea555984fcd2c6bf9e39d0f47a01b09e7c48)

x-pack/plugin/sql/qa/src/main/resources/agg.csv-spec

@@ -246,6 +246,194 @@ TRUNCATE(YEAR("birth_date"), -2)
 null
 1900
 ;
+// Fails for H2
+groupByCastScalarWithNumericRef
+SELECT CAST(ABS(EXTRACT(YEAR FROM "birth_date")) AS BIGINT) FROM test_emp GROUP BY 1 ORDER BY 1 NULLS FIRST;
+
+CAST(ABS(EXTRACT(YEAR FROM "birth_date")) AS BIGINT):l
+------------------------------------------------------
+null
+1952
+1953
+1954
+1955
+1956
+1957
+1958
+1959
+1960
+1961
+1962
+1963
+1964
+1965
+;
+
+groupByConvertScalar
+SELECT CONVERT(ABS(EXTRACT(YEAR FROM "birth_date")), SQL_BIGINT) FROM test_emp GROUP BY CONVERT(ABS(EXTRACT(YEAR FROM "birth_date")), SQL_BIGINT) ORDER BY CONVERT(ABS(EXTRACT(YEAR FROM "birth_date")), SQL_BIGINT) NULLS FIRST;
+
+
+CONVERT(ABS(EXTRACT(YEAR FROM "birth_date")), SQL_BIGINT):l
+-----------------------------------------------------------
+null
+1952
+1953
+1954
+1955
+1956
+1957
+1958
+1959
+1960
+1961
+1962
+1963
+1964
+1965
+;
+
+
+groupByConvertScalarWithAlias
+SELECT CONVERT(ABS(EXTRACT(YEAR FROM "birth_date")), SQL_BIGINT) as "convert" FROM test_emp GROUP BY "convert" ORDER BY "convert" NULLS FIRST;
+
+convert:l
+---------
+null
+1952
+1953
+1954
+1955
+1956
+1957
+1958
+1959
+1960
+1961
+1962
+1963
+1964
+1965
+;
+
+groupByConvertScalarWithNumericRef
+SELECT CONVERT(ABS(EXTRACT(YEAR FROM "birth_date")), SQL_BIGINT) FROM test_emp GROUP BY 1 ORDER BY 1 NULLS FIRST;
+
+CONVERT(ABS(EXTRACT(YEAR FROM "birth_date")), SQL_BIGINT):l
+-----------------------------------------------------------
+null
+1952
+1953
+1954
+1955
+1956
+1957
+1958
+1959
+1960
+1961
+1962
+1963
+1964
+1965
+;
+
+groupByConstantScalar
+SELECT PI() * emp_no FROM test_emp GROUP BY PI() * emp_no ORDER BY PI() * emp_no LIMIT 10;
+
+PI() * emp_no:d
+---------------
+31419.0681285515
+31422.2097212051
+31425.3513138587
+31428.4929065123
+31431.6344991659
+31434.7760918195
+31437.9176844731
+31441.0592771266
+31444.2008697802
+31447.3424624338
+;
+
+groupByConstantScalarWithOrderByDesc
+SELECT PI() * emp_no FROM test_emp GROUP BY PI() * emp_no ORDER BY PI() * emp_no DESC LIMIT 10;
+
+PI() * emp_no:d
+-------
+31730.0858012569
+31726.9442086033
+31723.8026159497
+31720.6610232961
+31717.5194306425
+31714.3778379889
+31711.2362453353
+31708.0946526817
+31704.9530600281
+31701.8114673746
+;
+
+groupByConstantScalarWithAlias
+SELECT PI() * emp_no AS "value" FROM test_emp GROUP BY value ORDER BY value LIMIT 10;
+
+value:d
+-------
+31419.0681285515
+31422.2097212051
+31425.3513138587
+31428.4929065123
+31431.6344991659
+31434.7760918195
+31437.9176844731
+31441.0592771266
+31444.2008697802
+31447.3424624338
+;
+
+groupByConstantScalarWithNumericRef
+SELECT PI() * emp_no FROM test_emp GROUP BY 1 ORDER BY 1 DESC LIMIT 10;
+
+PI() * emp_no:d
+-------
+31730.0858012569
+31726.9442086033
+31723.8026159497
+31720.6610232961
+31717.5194306425
+31714.3778379889
+31711.2362453353
+31708.0946526817
+31704.9530600281
+31701.8114673746
+;
+
+groupByFieldAndConstantScalarWithMultipleOrderBy
+SELECT gender, emp_no % 3 + PI() FROM test_emp GROUP BY gender, emp_no % 3 + PI() ORDER BY gender, emp_no % 3 + PI() DESC LIMIT 8;
+
+gender:s    |emp_no % 3 + PI():d
+------------+------------------
+null        |5.1415926535
+null        |4.1415926535
+null        |3.1415926535
+F           |5.1415926535
+F           |4.1415926535
+F           |3.1415926535
+M           |5.1415926535
+M           |4.1415926535
+;
+
+groupByFieldAndConstantScalarWithAliasWithOrderByDesc
+SELECT gender, emp_no % 3 + PI() as p FROM test_emp GROUP BY gender, emp_no % 3 + PI() ORDER BY gender DESC, p DESC LIMIT 8;
+
+gender:s    |p:d
+------------+------------------
+M           |5.1415926535
+M           |4.1415926535
+M           |3.1415926535
+F           |5.1415926535
+F           |4.1415926535
+F           |3.1415926535
+null        |5.1415926535
+null        |4.1415926535
+;
 
 //
 // Grouping functions

x-pack/plugin/sql/qa/src/main/resources/agg.sql-spec

@@ -51,6 +51,10 @@ groupByMulScalar
 SELECT emp_no * 2 AS e FROM test_emp GROUP BY e ORDER BY e;
 groupByModScalar
 SELECT (emp_no % 3) + 1 AS e FROM test_emp GROUP BY e ORDER BY e;
+groupByCastScalar
+SELECT CAST(ABS(EXTRACT(YEAR FROM "birth_date")) AS BIGINT) FROM test_emp GROUP BY CAST(ABS(EXTRACT(YEAR FROM "birth_date")) AS BIGINT) ORDER BY CAST(ABS(EXTRACT(YEAR FROM "birth_date")) AS BIGINT) NULLS FIRST;
+groupByCastScalarWithAlias
+SELECT CAST(ABS(EXTRACT(YEAR FROM "birth_date")) AS BIGINT) as "cast" FROM test_emp GROUP BY "cast" ORDER BY "cast" NULLS FIRST;
 
 // group by nested functions with no alias
 groupByTruncate

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/analysis/analyzer/Analyzer.java

@@ -66,6 +66,7 @@ import java.util.List;
 import java.util.Map;
 import java.util.Objects;
 import java.util.Set;
+import java.util.stream.Collectors;
 
 import static java.util.Collections.emptyList;
 import static java.util.Collections.singletonList;
@@ -609,12 +610,15 @@ public class Analyzer extends RuleExecutor<LogicalPlan> {
                         .map(or -> tryResolveExpression(or, o.child()))
                         .collect(toList());
 
-                AttributeSet resolvedRefs = Expressions.references(maybeResolved.stream()
+
+                Set<Expression> resolvedRefs = maybeResolved.stream()
                     .filter(Expression::resolved)
-                        .collect(toList()));
+                    .collect(Collectors.toSet());
 
-
+                AttributeSet missing = Expressions.filterReferences(
-                AttributeSet missing = resolvedRefs.subtract(o.child().outputSet());
+                    resolvedRefs,
+                    o.child().outputSet()
+                );
 
                 if (!missing.isEmpty()) {
                     // Add missing attributes but project them away afterwards

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/analysis/analyzer/VerificationException.java

@@ -27,7 +27,7 @@ public class VerificationException extends AnalysisException {
     public String getMessage() {
         return failures.stream()
                 .map(f -> {
-                    Location l = f.source().source().source();
+                    Location l = f.node().source().source();
                     return "line " + l.getLineNumber() + ":" + l.getColumnNumber() + ": " + f.message();
                 })
                 .collect(Collectors.joining(StringUtils.NEW_LINE, "Found " + failures.size() + " problem(s)\n", StringUtils.EMPTY));

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/analysis/analyzer/Verifier.java

@@ -84,16 +84,16 @@ public final class Verifier {
     }
 
     static class Failure {
-        private final Node<?> source;
+        private final Node<?> node;
         private final String message;
 
-        Failure(Node<?> source, String message) {
+        Failure(Node<?> node, String message) {
-            this.source = source;
+            this.node = node;
             this.message = message;
         }
 
-        Node<?> source() {
+        Node<?> node() {
-            return source;
+            return node;
         }
 
         String message() {
@@ -102,7 +102,7 @@ public final class Verifier {
 
         @Override
         public int hashCode() {
-            return source.hashCode();
+            return Objects.hash(node);
         }
 
         @Override
@@ -116,7 +116,7 @@ public final class Verifier {
             }
 
             Verifier.Failure other = (Verifier.Failure) obj;
-            return Objects.equals(source, other.source);
+            return Objects.equals(node, other.node);
         }
 
         @Override
@@ -131,7 +131,7 @@ public final class Verifier {
 
     public Map<Node<?>, String> verifyFailures(LogicalPlan plan) {
         Collection<Failure> failures = verify(plan);
-        return failures.stream().collect(toMap(Failure::source, Failure::message));
+        return failures.stream().collect(toMap(Failure::node, Failure::message));
     }
 
     Collection<Failure> verify(LogicalPlan plan) {

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/AttributeMap.java

@@ -32,14 +32,14 @@ public class AttributeMap<E> implements Map<Attribute, E> {
 
         @Override
         public int hashCode() {
-            return attr.semanticHash();
+            return attr.hashCode();
         }
 
         @Override
         public boolean equals(Object obj) {
             if (obj instanceof AttributeWrapper) {
                 AttributeWrapper aw = (AttributeWrapper) obj;
-                return attr.semanticEquals(aw.attr);
+                return attr.equals(aw.attr);
             }
 
             return false;

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/Expression.java

@@ -126,9 +126,6 @@ public abstract class Expression extends Node<Expression> implements Resolvable
 
     public abstract DataType dataType();
 
-    @Override
-    public abstract int hashCode();
-
     @Override
     public String toString() {
         return nodeName() + "[" + propertiesToString(false) + "]";

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/Expressions.java

@@ -16,6 +16,7 @@ import java.util.LinkedHashSet;
 import java.util.List;
 import java.util.Set;
 import java.util.function.Predicate;
+import java.util.stream.Collectors;
 
 import static java.util.Collections.emptyList;
 import static java.util.Collections.emptyMap;
@@ -102,6 +103,31 @@ public final class Expressions {
         return set;
     }
 
+    public static AttributeSet filterReferences(Set<? extends Expression> exps, AttributeSet excluded) {
+        AttributeSet ret = new AttributeSet();
+        while (exps.size() > 0) {
+
+            Set<Expression> filteredExps = new LinkedHashSet<>();
+            for (Expression exp : exps) {
+                Expression attr = Expressions.attribute(exp);
+                if (attr == null || (excluded.contains(attr) == false)) {
+                    filteredExps.add(exp);
+                }
+            }
+
+            ret.addAll(new AttributeSet(
+                filteredExps.stream().filter(c->c.children().isEmpty())
+                .flatMap(exp->exp.references().stream())
+                .collect(Collectors.toSet())
+            ));
+
+            exps = filteredExps.stream()
+                .flatMap((Expression exp)->exp.children().stream())
+                .collect(Collectors.toSet());
+        }
+        return ret;
+    }
+
     public static String name(Expression e) {
         return e instanceof NamedExpression ? ((NamedExpression) e).name() : e.nodeName();
     }

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/FieldAttribute.java

@@ -102,11 +102,6 @@ public class FieldAttribute extends TypedAttribute {
         return new FieldAttribute(source(), this, name() + "." + type.getName(), type, qualifier(), nullable(), id(), synthetic());
     }
 
-    @Override
-    protected Expression canonicalize() {
-        return new FieldAttribute(source(), null, "<none>", field, null, Nullability.TRUE, id(), false);
-    }
-
     @Override
     protected Attribute clone(Source source, String name, DataType type, String qualifier,
             Nullability nullability, ExpressionId id, boolean synthetic) {

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/NamedExpression.java

@@ -67,7 +67,7 @@ public abstract class NamedExpression extends Expression {
 
     @Override
     public int hashCode() {
-        return Objects.hash(id, name, synthetic);
+        return Objects.hash(super.hashCode(), name, synthetic);
     }
 
     @Override
@@ -81,7 +81,6 @@ public abstract class NamedExpression extends Expression {
 
         NamedExpression other = (NamedExpression) obj;
         return Objects.equals(synthetic, other.synthetic)
-                && Objects.equals(id, other.id)
                 /*
                  * It is important that the line below be `name`
                  * and not `name()` because subclasses might override

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/function/FunctionAttribute.java

@@ -29,11 +29,11 @@ public abstract class FunctionAttribute extends TypedAttribute {
 
     @Override
     public int hashCode() {
-        return Objects.hash(super.hashCode(), functionId);
+        return Objects.hash(super.hashCode());
     }
 
     @Override
     public boolean equals(Object obj) {
-        return super.equals(obj) && Objects.equals(functionId, ((FunctionAttribute) obj).functionId());
+        return super.equals(obj);
     }
 }

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/function/aggregate/AggregateFunctionAttribute.java

@@ -75,14 +75,14 @@ public class AggregateFunctionAttribute extends FunctionAttribute {
 
     @Override
     public int hashCode() {
-        return Objects.hash(super.hashCode(), innerId, propertyPath);
+        return Objects.hash(super.hashCode(), propertyPath);
     }
 
     @Override
     public boolean equals(Object obj) {
         if (super.equals(obj)) {
             AggregateFunctionAttribute other = (AggregateFunctionAttribute) obj;
-            return Objects.equals(innerId, other.innerId) && Objects.equals(propertyPath, other.propertyPath);
+            return Objects.equals(propertyPath, other.propertyPath);
         }
         return false;
     }

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/function/aggregate/Count.java

@@ -78,11 +78,15 @@ public class Count extends AggregateFunction {
     
     @Override
     public boolean equals(Object obj) {
-        if (false == super.equals(obj)) {
+        if (this == obj) {
+            return true;
+        }
+        if (obj == null || obj.getClass() != getClass()) {
             return false;
         }
         Count other = (Count) obj;
-        return Objects.equals(other.distinct(), distinct());
+        return Objects.equals(other.distinct(), distinct())
+            && Objects.equals(field(), other.field());
     }
 
     @Override

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/gen/script/Param.java

@@ -5,6 +5,8 @@
  */
 package org.elasticsearch.xpack.sql.expression.gen.script;
 
+import java.util.Objects;
+
 import static org.elasticsearch.common.logging.LoggerMessageFormat.format;
 
 abstract class Param<T> {
@@ -24,4 +26,24 @@ abstract class Param<T> {
     public String toString() {
         return format(null, "{{}={}}", prefix(), value);
     }
+
+
+    @Override
+    public int hashCode() {
+        if (this.value == null) {
+            return Objects.hashCode(null);
+        }
+        return this.value.hashCode();
+    }
+
+    @Override
+    public boolean equals(Object obj) {
+        if ((obj instanceof  Param) == false) {
+            return false;
+        }
+        if (this.value == null) {
+            return ((Param)obj).value == null;
+        }
+        return this.value.equals(((Param)obj).value);
+    }
 }

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/expression/gen/script/Params.java

@@ -124,4 +124,17 @@ public class Params {
     public String toString() {
         return params.toString();
     }
+
+    @Override
+    public int hashCode() {
+        return this.params.hashCode();
+    }
+
+    @Override
+    public boolean equals(Object obj) {
+        if ((obj instanceof  Params) == false) {
+            return false;
+        }
+        return this.params.equals(((Params)obj).params);
+    }
 }

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/plan/logical/Pivot.java

@@ -20,6 +20,7 @@ import org.elasticsearch.xpack.sql.tree.Source;
 import java.util.ArrayList;
 import java.util.List;
 import java.util.Objects;
+import java.util.stream.Collectors;
 
 import static java.util.Collections.singletonList;
 
@@ -40,6 +41,14 @@ public class Pivot extends UnaryPlan {
         this.aggregates = aggregates;
     }
 
+    private static Expression withQualifierNull(Expression e) {
+        if (e instanceof Attribute) {
+            Attribute fa = (Attribute) e;
+            return fa.withQualifier(null);
+        }
+        return e;
+    }
+
     @Override
     protected NodeInfo<Pivot> info() {
         return NodeInfo.create(this, Pivot::new, child(), column, values, aggregates);
@@ -47,7 +56,22 @@ public class Pivot extends UnaryPlan {
 
     @Override
     protected Pivot replaceChild(LogicalPlan newChild) {
-        return new Pivot(source(), newChild, column, values, aggregates);
+        Expression newColumn = column;
+        List<NamedExpression> newAggregates = aggregates;
+
+        if (newChild instanceof EsRelation) {
+            // when changing from a SubQueryAlias to EsRelation
+            // the qualifier of the column and aggregates needs
+            // to be changed to null like the attributes of EsRelation
+            // otherwise they don't equal and aren't removed
+            // when calculating the groupingSet
+            newColumn = column.transformUp(Pivot::withQualifierNull);
+            newAggregates = aggregates.stream().map((NamedExpression aggregate) ->
+                (NamedExpression) aggregate.transformUp(Pivot::withQualifierNull)
+            ).collect(Collectors.toList());
+        }
+
+        return new Pivot(source(), newChild, newColumn, values, newAggregates);
     }
 
     public Expression column() {
@@ -83,7 +107,7 @@ public class Pivot extends UnaryPlan {
             if (aggregates.size() == 1) {
                 NamedExpression agg = aggregates.get(0);
                 for (NamedExpression value : values) {
-                    ExpressionId id = new ExpressionId(agg.id().hashCode() + value.id().hashCode());
+                    ExpressionId id = value.id();
                     out.add(value.toAttribute().withDataType(agg.dataType()).withId(id));
                 }
             }
@@ -92,7 +116,7 @@ public class Pivot extends UnaryPlan {
                 for (NamedExpression agg : aggregates) {
                     String name = agg instanceof Function ? ((Function) agg).functionName() : agg.name();
                     for (NamedExpression value : values) {
-                        ExpressionId id = new ExpressionId(agg.id().hashCode() + value.id().hashCode());
+                        ExpressionId id = value.id();
                         out.add(value.toAttribute().withName(value.name() + "_" + name).withDataType(agg.dataType()).withId(id));
                     }
                 }

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

@@ -72,6 +72,7 @@ import org.elasticsearch.xpack.sql.util.Check;
 
 import java.util.ArrayList;
 import java.util.Arrays;
+import java.util.HashMap;
 import java.util.LinkedHashMap;
 import java.util.List;
 import java.util.Map;
@@ -122,7 +123,7 @@ class QueryFolder extends RuleExecutor<PhysicalPlan> {
                 EsQueryExec exec = (EsQueryExec) project.child();
                 QueryContainer queryC = exec.queryContainer();
 
-                Map<Attribute, Attribute> aliases = new LinkedHashMap<>(queryC.aliases());
+                Map<ExpressionId, Attribute> aliases = new LinkedHashMap<>(queryC.aliases());
                 Map<Attribute, Pipe> processors = new LinkedHashMap<>(queryC.scalarFunctions());
 
                 for (NamedExpression pj : project.projections()) {
@@ -132,7 +133,7 @@ class QueryFolder extends RuleExecutor<PhysicalPlan> {
 
                         if (e instanceof NamedExpression) {
                             Attribute attr = ((NamedExpression) e).toAttribute();
-                            aliases.put(aliasAttr, attr);
+                            aliases.put(aliasAttr.id(), attr);
                             // add placeholder for each scalar function
                             if (e instanceof ScalarFunction) {
                                 processors.put(attr, Expressions.pipe(e));
@@ -153,7 +154,7 @@ class QueryFolder extends RuleExecutor<PhysicalPlan> {
                 }
 
                 QueryContainer clone = new QueryContainer(queryC.query(), queryC.aggs(), queryC.fields(),
-                        new AttributeMap<>(aliases),
+                        new HashMap<>(aliases),
                         queryC.pseudoFunctions(),
                         new AttributeMap<>(processors),
                         queryC.sort(),
@@ -234,7 +235,7 @@ class QueryFolder extends RuleExecutor<PhysicalPlan> {
                 queryC = queryC.addGroups(groupingContext.groupMap.values());
             }
 
-            Map<Attribute, Attribute> aliases = new LinkedHashMap<>();
+            Map<ExpressionId, Attribute> aliases = new LinkedHashMap<>();
             // tracker for compound aggs seen in a group
             Map<CompoundNumericAggregate, String> compoundAggMap = new LinkedHashMap<>();
 
@@ -262,7 +263,7 @@ class QueryFolder extends RuleExecutor<PhysicalPlan> {
 
                     // record aliases in case they are later referred in the tree
                     if (as != null && as.child() instanceof NamedExpression) {
-                        aliases.put(as.toAttribute(), ((NamedExpression) as.child()).toAttribute());
+                        aliases.put(as.toAttribute().id(), ((NamedExpression) as.child()).toAttribute());
                     }
 
                     //
@@ -392,9 +393,9 @@ class QueryFolder extends RuleExecutor<PhysicalPlan> {
                 }
 
             if (!aliases.isEmpty()) {
-                Map<Attribute, Attribute> newAliases = new LinkedHashMap<>(queryC.aliases());
+                Map<ExpressionId, Attribute> newAliases = new LinkedHashMap<>(queryC.aliases());
                 newAliases.putAll(aliases);
-                queryC = queryC.withAliases(new AttributeMap<>(newAliases));
+                queryC = queryC.withAliases(new HashMap<>(newAliases));
             }
             return new EsQueryExec(exec.source(), exec.index(), a.output(), queryC);
         }
@@ -481,21 +482,13 @@ class QueryFolder extends RuleExecutor<PhysicalPlan> {
                     // check whether sorting is on an group (and thus nested agg) or field
                     Attribute attr = ((NamedExpression) order.child()).toAttribute();
                     // check whether there's an alias (occurs with scalar functions which are not named)
-                    attr = qContainer.aliases().getOrDefault(attr, attr);
+                    attr = qContainer.aliases().getOrDefault(attr.id(), attr);
-                    String lookup = attr.id().toString();
+                    GroupByKey group = qContainer.findGroupForAgg(attr);
-                    GroupByKey group = qContainer.findGroupForAgg(lookup);
 
                     // TODO: might need to validate whether the target field or group actually exist
                     if (group != null && group != Aggs.IMPLICIT_GROUP_KEY) {
-                        // check whether the lookup matches a group
-                        if (group.id().equals(lookup)) {
                         qContainer = qContainer.updateGroup(group.with(direction));
                     }
-                        // else it's a leafAgg
-                        else {
-                            qContainer = qContainer.updateGroup(group.with(direction));
-                        }
-                    }
                     else {
                         // scalar functions typically require script ordering
                         if (attr instanceof ScalarFunctionAttribute) {
@@ -504,7 +497,7 @@ class QueryFolder extends RuleExecutor<PhysicalPlan> {
                             if (sfa.orderBy() != null) {
                                 if (sfa.orderBy() instanceof NamedExpression) {
                                     Attribute at = ((NamedExpression) sfa.orderBy()).toAttribute();
-                                    at = qContainer.aliases().getOrDefault(at, at);
+                                    at = qContainer.aliases().getOrDefault(at.id(), at);
                                     qContainer = qContainer.addSort(new AttributeSort(at, direction, missing));
                                 } else if (!sfa.orderBy().foldable()) {
                                     // ignore constant

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

@@ -12,7 +12,6 @@ import org.elasticsearch.search.sort.SortOrder;
 import org.elasticsearch.xpack.sql.SqlIllegalArgumentException;
 import org.elasticsearch.xpack.sql.expression.Attribute;
 import org.elasticsearch.xpack.sql.expression.Expression;
-import org.elasticsearch.xpack.sql.expression.ExpressionId;
 import org.elasticsearch.xpack.sql.expression.Expressions;
 import org.elasticsearch.xpack.sql.expression.FieldAttribute;
 import org.elasticsearch.xpack.sql.expression.Foldables;
@@ -210,14 +209,14 @@ final class QueryTranslator {
     }
 
     static class GroupingContext {
-        final Map<ExpressionId, GroupByKey> groupMap;
+        final Map<NamedExpression, GroupByKey> groupMap;
         final GroupByKey tail;
 
-        GroupingContext(Map<ExpressionId, GroupByKey> groupMap) {
+        GroupingContext(Map<NamedExpression, GroupByKey> groupMap) {
             this.groupMap = groupMap;
 
             GroupByKey lastAgg = null;
-            for (Entry<ExpressionId, GroupByKey> entry : groupMap.entrySet()) {
+            for (Entry<NamedExpression, GroupByKey> entry : groupMap.entrySet()) {
                 lastAgg = entry.getValue();
             }
 
@@ -232,7 +231,7 @@ final class QueryTranslator {
                     GroupByKey matchingGroup = null;
                     // group found - finding the dedicated agg
                     if (f.field() instanceof NamedExpression) {
-                        matchingGroup = groupMap.get(((NamedExpression) f.field()).id());
+                        matchingGroup = groupMap.get(f.field());
                     }
                     // return matching group or the tail (last group)
                     return matchingGroup != null ? matchingGroup : tail;
@@ -242,7 +241,7 @@ final class QueryTranslator {
                 }
             }
             if (exp instanceof NamedExpression) {
-                return groupMap.get(((NamedExpression) exp).id());
+                return groupMap.get(exp);
             }
             throw new SqlIllegalArgumentException("Don't know how to find group for expression {}", exp);
         }
@@ -261,18 +260,18 @@ final class QueryTranslator {
             return null;
         }
 
-        Map<ExpressionId, GroupByKey> aggMap = new LinkedHashMap<>();
+        Map<NamedExpression, GroupByKey> aggMap = new LinkedHashMap<>();
 
         for (Expression exp : groupings) {
             GroupByKey key = null;
-            ExpressionId id;
+            NamedExpression id;
             String aggId;
 
             if (exp instanceof NamedExpression) {
                 NamedExpression ne = (NamedExpression) exp;
 
-                id = ne.id();
+                id = ne;
-                aggId = id.toString();
+                aggId = ne.id().toString();
 
                 // change analyzed to non non-analyzed attributes
                 if (exp instanceof FieldAttribute) {

x-pack/plugin/sql/src/main/java/org/elasticsearch/xpack/sql/querydsl/agg/Aggs.java

@@ -10,6 +10,8 @@ import org.elasticsearch.search.aggregations.bucket.composite.CompositeAggregati
 import org.elasticsearch.search.aggregations.bucket.composite.CompositeValuesSourceBuilder;
 import org.elasticsearch.search.aggregations.bucket.filter.FiltersAggregationBuilder;
 import org.elasticsearch.xpack.sql.SqlIllegalArgumentException;
+import org.elasticsearch.xpack.sql.expression.Attribute;
+import org.elasticsearch.xpack.sql.expression.function.scalar.ScalarFunctionAttribute;
 import org.elasticsearch.xpack.sql.expression.gen.script.ScriptTemplate;
 import org.elasticsearch.xpack.sql.querydsl.container.Sort.Direction;
 import org.elasticsearch.xpack.sql.util.StringUtils;
@@ -121,16 +123,23 @@ public class Aggs {
         return new Aggs(groups, simpleAggs, combine(pipelineAggs, pipelineAgg));
     }
 
-    public GroupByKey findGroupForAgg(String groupOrAggId) {
+    public GroupByKey findGroupForAgg(Attribute attr) {
+        String id = attr.id().toString();
         for (GroupByKey group : this.groups) {
-            if (groupOrAggId.equals(group.id())) {
+            if (id.equals(group.id())) {
                 return group;
             }
+            if (attr instanceof ScalarFunctionAttribute) {
+                ScalarFunctionAttribute sfa = (ScalarFunctionAttribute) attr;
+                if (group.script() != null && group.script().equals(sfa.script())) {
+                    return group;
+                }
+            }
         }
 
         // maybe it's the default group agg ?
         for (Agg agg : simpleAggs) {
-            if (groupOrAggId.equals(agg.id())) {
+            if (id.equals(agg.id())) {
                 return IMPLICIT_GROUP_KEY;
             }
         }

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

@@ -38,6 +38,7 @@ import java.util.AbstractMap;
 import java.util.ArrayList;
 import java.util.BitSet;
 import java.util.Collection;
+import java.util.Collections;
 import java.util.Comparator;
 import java.util.LinkedHashMap;
 import java.util.LinkedHashSet;
@@ -70,7 +71,7 @@ public class QueryContainer {
     private final List<Tuple<FieldExtraction, ExpressionId>> fields;
 
     // aliases (maps an alias to its actual resolved attribute)
-    private final AttributeMap<Attribute> aliases;
+    private final Map<ExpressionId, Attribute> aliases;
 
     // pseudo functions (like count) - that are 'extracted' from other aggs
     private final Map<String, GroupByKey> pseudoFunctions;
@@ -98,7 +99,7 @@ public class QueryContainer {
             Aggs aggs,
             List<Tuple<FieldExtraction,
             ExpressionId>> fields,
-            AttributeMap<Attribute> aliases,
+            Map<ExpressionId, Attribute> aliases,
             Map<String, GroupByKey> pseudoFunctions,
             AttributeMap<Pipe> scalarFunctions,
             Set<Sort> sort,
@@ -109,7 +110,7 @@ public class QueryContainer {
         this.query = query;
         this.aggs = aggs == null ? Aggs.EMPTY : aggs;
         this.fields = fields == null || fields.isEmpty() ? emptyList() : fields;
-        this.aliases = aliases == null || aliases.isEmpty() ? AttributeMap.emptyAttributeMap() : aliases;
+        this.aliases = aliases == null || aliases.isEmpty() ? Collections.emptyMap() : aliases;
         this.pseudoFunctions = pseudoFunctions == null || pseudoFunctions.isEmpty() ? emptyMap() : pseudoFunctions;
         this.scalarFunctions = scalarFunctions == null || scalarFunctions.isEmpty() ? AttributeMap.emptyAttributeMap() : scalarFunctions;
         this.sort = sort == null || sort.isEmpty() ? emptySet() : sort;
@@ -141,7 +142,7 @@ public class QueryContainer {
                 if (as.attribute() instanceof AggregateFunctionAttribute) {
                     aggSort = true;
                     AggregateFunctionAttribute afa = (AggregateFunctionAttribute) as.attribute();
-                    afa = (AggregateFunctionAttribute) aliases.getOrDefault(afa, afa);
+                    afa = (AggregateFunctionAttribute) aliases.getOrDefault(afa.innerId(), afa);
                     int atIndex = -1;
                     for (int i = 0; i < fields.size(); i++) {
                         Tuple<FieldExtraction, ExpressionId> field = fields.get(i);
@@ -179,7 +180,7 @@ public class QueryContainer {
     public BitSet columnMask(List<Attribute> columns) {
         BitSet mask = new BitSet(fields.size());
         for (Attribute column : columns) {
-            Attribute alias = aliases.get(column);
+            Attribute alias = aliases.get(column.id());
             // find the column index
             int index = -1;
 
@@ -217,7 +218,7 @@ public class QueryContainer {
         return fields;
     }
 
-    public AttributeMap<Attribute> aliases() {
+    public Map<ExpressionId, Attribute> aliases() {
         return aliases;
     }
 
@@ -271,7 +272,7 @@ public class QueryContainer {
                 minPageSize);
     }
 
-    public QueryContainer withAliases(AttributeMap<Attribute> a) {
+    public QueryContainer withAliases(Map<ExpressionId, Attribute> a) {
         return new QueryContainer(query, aggs, fields, a, pseudoFunctions, scalarFunctions, sort, limit, trackHits, includeFrozen,
                 minPageSize);
     }
@@ -312,7 +313,7 @@ public class QueryContainer {
     }
 
     private String aliasName(Attribute attr) {
-        return aliases.getOrDefault(attr, attr).name();
+        return aliases.getOrDefault(attr.id(), attr).name();
     }
 
     //
@@ -397,7 +398,7 @@ public class QueryContainer {
 
     // replace function/operators's input with references
     private Tuple<QueryContainer, FieldExtraction> resolvedTreeComputingRef(ScalarFunctionAttribute ta) {
-        Attribute attribute = aliases.getOrDefault(ta, ta);
+        Attribute attribute = aliases.getOrDefault(ta.id(), ta);
         Pipe proc = scalarFunctions.get(attribute);
 
         // check the attribute itself
@@ -419,7 +420,7 @@ public class QueryContainer {
 
             @Override
             public FieldExtraction resolve(Attribute attribute) {
-                Attribute attr = aliases.getOrDefault(attribute, attribute);
+                Attribute attr = aliases.getOrDefault(attribute.id(), attribute);
                 Tuple<QueryContainer, FieldExtraction> ref = container.toReference(attr);
                 container = ref.v1();
                 return ref.v2();
@@ -486,8 +487,8 @@ public class QueryContainer {
         return with(aggs.addGroups(values));
     }
 
-    public GroupByKey findGroupForAgg(String aggId) {
+    public GroupByKey findGroupForAgg(Attribute attr) {
-        return aggs.findGroupForAgg(aggId);
+        return aggs.findGroupForAgg(attr);
     }
 
     public QueryContainer updateGroup(GroupByKey group) {

x-pack/plugin/sql/src/test/java/org/elasticsearch/xpack/sql/analysis/analyzer/VerifierErrorMessagesTests.java

@@ -377,7 +377,8 @@ public class VerifierErrorMessagesTests extends ESTestCase {
     }
 
     public void testMultipleColumns() {
-        assertEquals("1:43: Unknown column [xxx]\nline 1:8: Unknown column [xxx]",
+        // We get only one message back because the messages are grouped by the node that caused the issue
+        assertEquals("1:43: Unknown column [xxx]",
                 error("SELECT xxx FROM test GROUP BY DAY_oF_YEAR(xxx)"));
     }
 

x-pack/plugin/sql/src/test/java/org/elasticsearch/xpack/sql/expression/AttributeMapTests.java

@@ -55,7 +55,7 @@ public class AttributeMapTests extends ESTestCase {
 
         Attribute one = m.keySet().iterator().next();
         assertThat(m.containsKey(one), is(true));
-        assertThat(m.containsKey(a("one")), is(false));
+        assertThat(m.containsKey(a("one")), is(true));
         assertThat(m.containsValue("one"), is(true));
         assertThat(m.containsValue("on"), is(false));
         assertThat(m.attributeNames(), contains("one", "two", "three"));
@@ -74,7 +74,7 @@ public class AttributeMapTests extends ESTestCase {
         assertThat(m.isEmpty(), is(false));
 
         assertThat(m.containsKey(one), is(true));
-        assertThat(m.containsKey(a("one")), is(false));
+        assertThat(m.containsKey(a("one")), is(true));
         assertThat(m.containsValue("one"), is(true));
         assertThat(m.containsValue("on"), is(false));
     }

x-pack/plugin/sql/src/test/java/org/elasticsearch/xpack/sql/expression/UnresolvedAttributeTests.java

@@ -54,8 +54,6 @@ public class UnresolvedAttributeTests extends AbstractNodeTestCase<UnresolvedAtt
             () -> new UnresolvedAttribute(a.source(), a.name(),
                     randomValueOtherThan(a.qualifier(), UnresolvedAttributeTests::randomQualifier),
                     a.id(), a.unresolvedMessage(), a.resolutionMetadata()),
-            () -> new UnresolvedAttribute(a.source(), a.name(), a.qualifier(),
-                    new ExpressionId(), a.unresolvedMessage(), a.resolutionMetadata()),
             () -> new UnresolvedAttribute(a.source(), a.name(), a.qualifier(), a.id(),
                     randomValueOtherThan(a.unresolvedMessage(), () -> randomUnresolvedMessage()),
                     a.resolutionMetadata()),

x-pack/plugin/sql/src/test/java/org/elasticsearch/xpack/sql/optimizer/OptimizerTests.java

@@ -236,7 +236,7 @@ public class OptimizerTests extends ESTestCase {
         assertTrue(result instanceof Project);
         List<? extends NamedExpression> projections = ((Project) result).projections();
         assertEquals(2, projections.size());
-        assertSame(projections.get(0), projections.get(1));
+        assertEquals(projections.get(0), projections.get(1));
     }
 
     public void testCombineProjections() {

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

@@ -1132,6 +1132,257 @@ public class QueryTranslatorTests extends ESTestCase {
                         + "\"gap_policy\":\"skip\"}}}}}"));
     }
 
+    public void testGroupByCastScalar() {
+        PhysicalPlan p = optimizeAndPlan("SELECT CAST(ABS(EXTRACT(YEAR FROM date)) AS BIGINT) FROM test " +
+            "GROUP BY CAST(ABS(EXTRACT(YEAR FROM date)) AS BIGINT) ORDER BY CAST(ABS(EXTRACT(YEAR FROM date)) AS BIGINT) NULLS FIRST");
+        assertEquals(EsQueryExec.class, p.getClass());
+        assertEquals(1, p.output().size());
+        assertEquals("CAST(ABS(EXTRACT(YEAR FROM date)) AS BIGINT)", p.output().get(0).qualifiedName());
+        assertEquals(DataType.LONG, p.output().get(0).dataType());
+        assertThat(
+            ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                .replaceAll("\\s+", ""),
+            endsWith("{\"source\":\"InternalSqlScriptUtils.cast(InternalSqlScriptUtils.abs(InternalSqlScriptUtils.dateTimeChrono" +
+                "(InternalSqlScriptUtils.docValue(doc,params.v0),params.v1,params.v2)),params.v3)\",\"lang\":\"painless\"," +
+                "\"params\":{\"v0\":\"date\",\"v1\":\"Z\",\"v2\":\"YEAR\",\"v3\":\"LONG\"}},\"missing_bucket\":true," +
+                "\"value_type\":\"long\",\"order\":\"asc\"}}}]}}}")
+        );
+    }
+
+    public void testGroupByCastScalarWithAlias() {
+        PhysicalPlan p = optimizeAndPlan("SELECT CAST(ABS(EXTRACT(YEAR FROM date)) AS BIGINT) as \"cast\"  FROM test " +
+            "GROUP BY \"cast\" ORDER BY \"cast\" NULLS FIRST");
+        assertEquals(EsQueryExec.class, p.getClass());
+        assertEquals(1, p.output().size());
+        assertEquals("cast", p.output().get(0).qualifiedName());
+        assertEquals(DataType.LONG, p.output().get(0).dataType());
+        assertThat(
+            ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                .replaceAll("\\s+", ""),
+            endsWith("{\"source\":\"InternalSqlScriptUtils.cast(InternalSqlScriptUtils.abs(InternalSqlScriptUtils.dateTimeChrono" +
+                "(InternalSqlScriptUtils.docValue(doc,params.v0),params.v1,params.v2)),params.v3)\",\"lang\":\"painless\"," +
+                "\"params\":{\"v0\":\"date\",\"v1\":\"Z\",\"v2\":\"YEAR\",\"v3\":\"LONG\"}},\"missing_bucket\":true," +
+                "\"value_type\":\"long\",\"order\":\"asc\"}}}]}}}")
+        );
+    }
+
+    public void testGroupByCastScalarWithNumericRef() {
+        PhysicalPlan p = optimizeAndPlan("SELECT CAST(ABS(EXTRACT(YEAR FROM date)) AS BIGINT) FROM test " +
+            "GROUP BY 1 ORDER BY 1 NULLS FIRST");
+        assertEquals(EsQueryExec.class, p.getClass());
+        assertEquals(1, p.output().size());
+        assertEquals("CAST(ABS(EXTRACT(YEAR FROM date)) AS BIGINT)", p.output().get(0).qualifiedName());
+        assertEquals(DataType.LONG, p.output().get(0).dataType());
+        assertThat(
+            ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                .replaceAll("\\s+", ""),
+            endsWith("{\"source\":\"InternalSqlScriptUtils.cast(InternalSqlScriptUtils.abs(InternalSqlScriptUtils.dateTimeChrono" +
+                "(InternalSqlScriptUtils.docValue(doc,params.v0),params.v1,params.v2)),params.v3)\",\"lang\":\"painless\"," +
+                "\"params\":{\"v0\":\"date\",\"v1\":\"Z\",\"v2\":\"YEAR\",\"v3\":\"LONG\"}},\"missing_bucket\":true," +
+                "\"value_type\":\"long\",\"order\":\"asc\"}}}]}}}")
+        );
+    }
+
+    public void testGroupByConvertScalar() {
+        {
+            PhysicalPlan p = optimizeAndPlan("SELECT CONVERT(ABS(EXTRACT(YEAR FROM date)), SQL_BIGINT) FROM test " +
+                "GROUP BY CONVERT(ABS(EXTRACT(YEAR FROM date)), SQL_BIGINT) ORDER BY CONVERT(ABS(EXTRACT(YEAR FROM date)), SQL_BIGINT) " +
+                "NULLS FIRST");
+            assertEquals(EsQueryExec.class, p.getClass());
+            assertEquals(1, p.output().size());
+            assertEquals("CONVERT(ABS(EXTRACT(YEAR FROM date)), SQL_BIGINT)", p.output().get(0).qualifiedName());
+            assertEquals(DataType.LONG, p.output().get(0).dataType());
+            assertThat(
+                ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                    .replaceAll("\\s+", ""),
+                endsWith("{\"source\":\"InternalSqlScriptUtils.cast(InternalSqlScriptUtils.abs(InternalSqlScriptUtils.dateTimeChrono" +
+                    "(InternalSqlScriptUtils.docValue(doc,params.v0),params.v1,params.v2)),params.v3)\",\"lang\":\"painless\"," +
+                    "\"params\":{\"v0\":\"date\",\"v1\":\"Z\",\"v2\":\"YEAR\",\"v3\":\"LONG\"}},\"missing_bucket\":true," +
+                    "\"value_type\":\"long\",\"order\":\"asc\"}}}]}}}")
+            );
+        }
+        {
+            PhysicalPlan p = optimizeAndPlan("SELECT EXTRACT(HOUR FROM CONVERT(date, SQL_TIMESTAMP)) FROM test GROUP BY " +
+                "EXTRACT(HOUR FROM CONVERT(date, SQL_TIMESTAMP))");
+            assertEquals(EsQueryExec.class, p.getClass());
+            assertEquals(1, p.output().size());
+            assertEquals("EXTRACT(HOUR FROM CONVERT(date, SQL_TIMESTAMP))", p.output().get(0).qualifiedName());
+            assertEquals(DataType.INTEGER, p.output().get(0).dataType());
+            assertThat(
+                ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                    .replaceAll("\\s+", ""),
+                endsWith("{\"source\":\"InternalSqlScriptUtils.dateTimeChrono(" +
+                    "InternalSqlScriptUtils.docValue(doc,params.v0),params.v1,params.v2)\",\"lang\":\"painless\"," +
+                    "\"params\":{\"v0\":\"date\",\"v1\":\"Z\",\"v2\":\"HOUR_OF_DAY\"}},\"missing_bucket\":true," +
+                    "\"value_type\":\"long\",\"order\":\"asc\"}}}]}}}")
+            );
+        }
+    }
+
+    public void testGroupByConvertScalarWithAlias() {
+        {
+            PhysicalPlan p = optimizeAndPlan("SELECT CONVERT(ABS(EXTRACT(YEAR FROM date)), SQL_BIGINT) as \"convert\" FROM test " +
+                "GROUP BY \"convert\" ORDER BY \"convert\" NULLS FIRST");
+            assertEquals(EsQueryExec.class, p.getClass());
+            assertEquals(1, p.output().size());
+            assertEquals("convert", p.output().get(0).qualifiedName());
+            assertEquals(DataType.LONG, p.output().get(0).dataType());
+            assertThat(
+                ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                    .replaceAll("\\s+", ""),
+                endsWith("{\"source\":\"InternalSqlScriptUtils.cast(InternalSqlScriptUtils.abs(InternalSqlScriptUtils.dateTimeChrono" +
+                    "(InternalSqlScriptUtils.docValue(doc,params.v0),params.v1,params.v2)),params.v3)\",\"lang\":\"painless\"," +
+                    "\"params\":{\"v0\":\"date\",\"v1\":\"Z\",\"v2\":\"YEAR\",\"v3\":\"LONG\"}},\"missing_bucket\":true," +
+                    "\"value_type\":\"long\",\"order\":\"asc\"}}}]}}}")
+            );
+        }
+        {
+            PhysicalPlan p = optimizeAndPlan("SELECT EXTRACT(MINUTE FROM CONVERT(date, SQL_TIMESTAMP)) x FROM test GROUP BY x");
+            assertEquals(EsQueryExec.class, p.getClass());
+            assertEquals(1, p.output().size());
+            assertEquals("x", p.output().get(0).qualifiedName());
+            assertEquals(DataType.INTEGER, p.output().get(0).dataType());
+            assertThat(
+                ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                    .replaceAll("\\s+", ""),
+                endsWith("{\"source\":\"InternalSqlScriptUtils.dateTimeChrono(" +
+                    "InternalSqlScriptUtils.docValue(doc,params.v0),params.v1,params.v2)\",\"lang\":\"painless\"," +
+                    "\"params\":{\"v0\":\"date\",\"v1\":\"Z\",\"v2\":\"MINUTE_OF_HOUR\"}}," +
+                    "\"missing_bucket\":true,\"value_type\":\"long\",\"order\":\"asc\"}}}]}}}")
+            );
+        }
+    }
+
+    public void testGroupByConvertScalarWithNumericRef() {
+        PhysicalPlan p = optimizeAndPlan("SELECT CONVERT(ABS(EXTRACT(YEAR FROM date)), SQL_BIGINT) FROM test " +
+            "GROUP BY 1 ORDER BY 1 NULLS FIRST");
+        assertEquals(EsQueryExec.class, p.getClass());
+        assertEquals(1, p.output().size());
+        assertEquals("CONVERT(ABS(EXTRACT(YEAR FROM date)), SQL_BIGINT)", p.output().get(0).qualifiedName());
+        assertEquals(DataType.LONG, p.output().get(0).dataType());
+        assertThat(
+            ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                .replaceAll("\\s+", ""),
+            endsWith("{\"source\":\"InternalSqlScriptUtils.cast(InternalSqlScriptUtils.abs(InternalSqlScriptUtils.dateTimeChrono" +
+                "(InternalSqlScriptUtils.docValue(doc,params.v0),params.v1,params.v2)),params.v3)\",\"lang\":\"painless\"," +
+                "\"params\":{\"v0\":\"date\",\"v1\":\"Z\",\"v2\":\"YEAR\",\"v3\":\"LONG\"}},\"missing_bucket\":true," +
+                "\"value_type\":\"long\",\"order\":\"asc\"}}}]}}}")
+        );
+    }
+
+    public void testGroupByConstantScalar() {
+        PhysicalPlan p = optimizeAndPlan("SELECT PI() * int FROM test WHERE PI() * int > 5.0 GROUP BY PI() * int " +
+            "ORDER BY PI() * int LIMIT 10");
+        assertEquals(EsQueryExec.class, p.getClass());
+        assertEquals(1, p.output().size());
+        assertEquals("PI() * int", p.output().get(0).qualifiedName());
+        assertEquals(DataType.DOUBLE, p.output().get(0).dataType());
+        assertThat(
+            ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                .replaceAll("\\s+", ""),
+            endsWith("{\"script\":{\"source\":\"InternalSqlScriptUtils.mul(params.v0,InternalSqlScriptUtils.docValue(doc,params.v1))\"," +
+                "\"lang\":\"painless\",\"params\":{\"v0\":3.141592653589793,\"v1\":\"int\"}},\"missing_bucket\":true," +
+                "\"value_type\":\"double\",\"order\":\"asc\"}}}]}}}")
+        );
+    }
+
+
+    public void testGroupByConstantScalarWithAlias() {
+        {
+            PhysicalPlan p = optimizeAndPlan("SELECT PI() * int AS \"value\"  FROM test GROUP BY \"value\" ORDER BY \"value\" LIMIT 10");
+            assertEquals(EsQueryExec.class, p.getClass());
+            assertEquals(1, p.output().size());
+            assertEquals("value", p.output().get(0).qualifiedName());
+            assertEquals(DataType.DOUBLE, p.output().get(0).dataType());
+            assertThat(
+                ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                    .replaceAll("\\s+", ""),
+                endsWith("{\"script\":{\"source\":\"InternalSqlScriptUtils.mul(params.v0,InternalSqlScriptUtils.docValue(doc,params.v1))" +
+                    "\",\"lang\":\"painless\",\"params\":{\"v0\":3.141592653589793,\"v1\":\"int\"}},\"missing_bucket\":true," +
+                    "\"value_type\":\"double\",\"order\":\"asc\"}}}]}}}")
+            );
+        }
+        {
+            PhysicalPlan p = optimizeAndPlan("select (3 < int) as multi_language, count(*) from test group by multi_language");
+            assertEquals(EsQueryExec.class, p.getClass());
+            assertEquals(2, p.output().size());
+            assertEquals("multi_language", p.output().get(0).qualifiedName());
+            assertEquals(DataType.BOOLEAN, p.output().get(0).dataType());
+            assertEquals("count(*)", p.output().get(1).qualifiedName());
+            assertEquals(DataType.LONG, p.output().get(1).dataType());
+            assertThat(
+                ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                    .replaceAll("\\s+", ""),
+                endsWith("{\"source\":\"InternalSqlScriptUtils.gt(InternalSqlScriptUtils.docValue(doc,params.v0),params.v1)\"," +
+                    "\"lang\":\"painless\",\"params\":{\"v0\":\"int\",\"v1\":3}}," +
+                    "\"missing_bucket\":true,\"value_type\":\"boolean\",\"order\":\"asc\"}}}]}}}")
+            );
+        }
+    }
+
+
+    public void testGroupByConstantScalarWithNumericRef() {
+        {
+            PhysicalPlan p = optimizeAndPlan("SELECT PI() * int FROM test GROUP BY 1 ORDER BY 1 LIMIT 10");
+            assertEquals(EsQueryExec.class, p.getClass());
+            assertEquals(1, p.output().size());
+            assertEquals("PI() * int", p.output().get(0).qualifiedName());
+            assertEquals(DataType.DOUBLE, p.output().get(0).dataType());
+            assertThat(
+                ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                    .replaceAll("\\s+", ""),
+                endsWith("{\"script\":{\"source\":\"InternalSqlScriptUtils.mul(params.v0,InternalSqlScriptUtils.docValue(doc,params.v1))" +
+                    "\",\"lang\":\"painless\",\"params\":{\"v0\":3.141592653589793,\"v1\":\"int\"}},\"missing_bucket\":true," +
+                    "\"value_type\":\"double\",\"order\":\"asc\"}}}]}}}")
+            );
+        }
+        {
+            PhysicalPlan p = optimizeAndPlan("SELECT PI() * int FROM test GROUP BY 1");
+            assertEquals(EsQueryExec.class, p.getClass());
+            assertEquals(1, p.output().size());
+            assertEquals("PI() * int", p.output().get(0).qualifiedName());
+            assertEquals(DataType.DOUBLE, p.output().get(0).dataType());
+            assertThat(
+                ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                    .replaceAll("\\s+", ""),
+                endsWith("{\"source\":\"InternalSqlScriptUtils.mul(params.v0,InternalSqlScriptUtils.docValue(doc,params.v1))\"," +
+                    "\"lang\":\"painless\",\"params\":{\"v0\":3.141592653589793,\"v1\":\"int\"}}," +
+                    "\"missing_bucket\":true,\"value_type\":\"double\",\"order\":\"asc\"}}}]}}}")
+            );
+        }
+        {
+            PhysicalPlan p = optimizeAndPlan("SELECT date + 1 * INTERVAL '1' DAY FROM test GROUP BY 1");
+            assertEquals(EsQueryExec.class, p.getClass());
+            assertEquals(1, p.output().size());
+            assertEquals("date + 1 * INTERVAL '1' DAY", p.output().get(0).qualifiedName());
+            assertEquals(DataType.DATETIME, p.output().get(0).dataType());
+            assertThat(
+                ((EsQueryExec) p).queryContainer().aggs().asAggBuilder().toString()
+                    .replaceAll("\\s+", ""),
+                endsWith("{\"source\":\"InternalSqlScriptUtils.add(InternalSqlScriptUtils.docValue(doc,params.v0)," +
+                    "InternalSqlScriptUtils.intervalDayTime(params.v1,params.v2))\"," +
+                    "\"lang\":\"painless\",\"params\":{\"v0\":\"date\",\"v1\":\"PT24H\",\"v2\":\"INTERVAL_DAY\"}}," +
+                    "\"missing_bucket\":true,\"value_type\":\"long\",\"order\":\"asc\"}}}]}}}")
+            );
+        }
+    }
+
+    public void testOrderByWithCastWithMissingRefs() {
+        PhysicalPlan p = optimizeAndPlan("SELECT keyword FROM test ORDER BY date::TIME, int LIMIT 5");
+        assertEquals(EsQueryExec.class, p.getClass());
+        assertEquals(1, p.output().size());
+        assertEquals("test.keyword", p.output().get(0).qualifiedName());
+        assertEquals(DataType.KEYWORD, p.output().get(0).dataType());
+        assertThat(
+            ((EsQueryExec) p).queryContainer().toString()
+                .replaceAll("\\s+", ""),
+            endsWith("\"sort\":[{\"_script\":{\"script\":{\"source\":\"InternalSqlScriptUtils.nullSafeSortString(InternalSqlScriptUtils" +
+                ".cast(InternalSqlScriptUtils.docValue(doc,params.v0),params.v1))\",\"lang\":\"painless\"," +
+                "\"params\":{\"v0\":\"date\",\"v1\":\"TIME\"}},\"type\":\"string\",\"order\":\"asc\"}},{\"int\":{\"order\":\"asc\"," +
+                "\"missing\":\"_last\",\"unmapped_type\":\"integer\"}}]}")
+        );
+    }
+
     public void testTopHitsAggregationWithOneArg() {
         {
             PhysicalPlan p = optimizeAndPlan("SELECT FIRST(keyword) FROM test");

x-pack/plugin/sql/src/test/java/org/elasticsearch/xpack/sql/querydsl/container/QueryContainerTests.java

@@ -8,7 +8,7 @@ package org.elasticsearch.xpack.sql.querydsl.container;
 import org.elasticsearch.test.ESTestCase;
 import org.elasticsearch.xpack.sql.expression.Alias;
 import org.elasticsearch.xpack.sql.expression.Attribute;
-import org.elasticsearch.xpack.sql.expression.AttributeMap;
+import org.elasticsearch.xpack.sql.expression.ExpressionId;
 import org.elasticsearch.xpack.sql.expression.FieldAttribute;
 import org.elasticsearch.xpack.sql.querydsl.query.BoolQuery;
 import org.elasticsearch.xpack.sql.querydsl.query.MatchAll;
@@ -81,11 +81,11 @@ public class QueryContainerTests extends ESTestCase {
         Attribute fourth = new FieldAttribute(Source.EMPTY, "fourth", esField);
         Alias firstAliased = new Alias(Source.EMPTY, "firstAliased", first);
 
-        Map<Attribute,Attribute> aliasesMap = new LinkedHashMap<>();
+        Map<ExpressionId,Attribute> aliasesMap = new LinkedHashMap<>();
-        aliasesMap.put(firstAliased.toAttribute(), first);
+        aliasesMap.put(firstAliased.id(), first);
 
         QueryContainer queryContainer = new QueryContainer()
-            .withAliases(new AttributeMap<>(aliasesMap))
+            .withAliases(aliasesMap)
             .addColumn(third)
             .addColumn(first)
             .addColumn(fourth)