Code formatting errors have been fixed.

Adjustments have been made to the README.
Fixed a bug with the <listagg overflow clause> behavior, now it is silently ignored.
The dsqlMatch function has been redesigned.

Redesigned behavior with DISTINCT. Multiple elements are now allowed in the ORDER BY.
I also added influences on the sorting direction.

Author: ChudaykinAlex

doc/sql.extensions/README.listagg

@@ -10,7 +10,7 @@ Authors:
 
 Format:
 	<listagg set function> ::=
-		LISTAGG <left paren> [ <set quantifier> ] <character value expression> <comma> <listagg separator> [ <listagg overflow clause> ] <right paren> <within group specification>
+		LISTAGG <left paren> [ <set quantifier> ] <character value expression> <comma> <listagg separator> [ <listagg overflow clause> ] <right paren> [ <within group specification> ]
 
 	<listagg separator> ::=
 		<character string literal>
@@ -36,14 +36,8 @@ Syntax Rules:
 
 	There is a <listagg overflow clause> rule in the standard, which is intended to output an error
 	when the output value overflows. Since the LIST function always returns a BLOB, it was decided
-	that this rule would be meaningless. It was not implemented and silently ignored if specified.
-
-	If DISTINCT is specified for LISTAGG, then ORDER BY <sort specification list> must fully match
-	<character value expression>
-
-Notes:
-	If DISTINCT is specified, the presence of WITHIN GROUP must obey the restriction and will not
-	affect the subsequent code execution.
+	that this rule would be meaningless. So the OVERFLOW clause is syntactically supported but
+	silently ignored if specified.
 
 Examples:
 CREATE TABLE TEST_T
@@ -68,6 +62,22 @@ SELECT LISTAGG (DISTINCT COL3, ':') FROM TEST_T;
 ====
 A:B
 
+SELECT LISTAGG (DISTINCT COL3, ':') WITHIN GROUP (ORDER BY COL3 ASCENDING) FROM TEST_T;
+====
+A:B
+
+SELECT LISTAGG (DISTINCT COL3, ':') WITHIN GROUP (ORDER BY COL3 DESCENDING) FROM TEST_T;
+====
+B:A
+
+SELECT LISTAGG (DISTINCT COL3, ':') WITHIN GROUP (ORDER BY COL3 DESCENDING, COL4, COL5) FROM TEST_T;
+====
+B:A
+
+SELECT LISTAGG (DISTINCT COL3, ':') WITHIN GROUP (ORDER BY COL4, COL3 DESCENDING, COL5) FROM TEST_T;
+====
+A:B
+
 SELECT LISTAGG (DISTINCT COL3, ':') WITHIN GROUP (ORDER BY COL2) FROM TEST_T;
 ====
 A:B
@@ -108,7 +118,7 @@ SELECT LISTAGG (COL2, ':') WITHIN GROUP (ORDER BY COL3 ASC, COL4 DESC) FROM TEST
 =======
 C:A:D:B
 
-SELECT LISTAGG (ALL COL6, ':')FROM TEST_T;
+SELECT LISTAGG (ALL COL6, ':') FROM TEST_T;
 =======
 П:Д:Ж:Й
 
@@ -120,6 +130,18 @@ SELECT LISTAGG (ALL COL2, ':') WITHIN GROUP (ORDER BY COL6) FROM TEST_T;
 =======
 B:C:D:A
 
+SELECT LISTAGG (COL4, ':' ON OVERFLOW TRUNCATE '...' WITHOUT COUNT) WITHIN GROUP (ORDER BY COL3 ASC) FROM TEST_T;
+=======
+J:L:I:K
+
+SELECT LISTAGG (COL4, ':' ON OVERFLOW TRUNCATE '...' WITH COUNT) WITHIN GROUP (ORDER BY COL3 DESC) FROM TEST_T;
+======
+I:K:J:L
+
+SELECT LISTAGG (DISTINCT COL3, ':' ON OVERFLOW ERROR) WITHIN GROUP (ORDER BY COL3) FROM TEST_T;
+===
+A:B
+
 INSERT INTO TEST_T values(5, 'E', NULL, NULL, NULL, NULL);
 INSERT INTO TEST_T values(6, 'F', 'C', 'N', true, 'К');
 
@@ -135,10 +157,3 @@ SELECT LISTAGG (ALL COL2, ':') WITHIN GROUP (ORDER BY COL6 NULLS FIRST) FROM TES
 ===========
 E:B:C:D:F:A
 
-SELECT LISTAGG (DISTINCT COL3, ':') WITHIN GROUP (ORDER BY COL2) FROM TEST_T;
-========
-Statement failed, SQLSTATE = 42000
-SQL error code = -104
--Invalid command
--Sort-key of the ORDER BY specification must match the argument list
-

src/common/ParserTokens.h

@@ -205,6 +205,7 @@ PARSER_TOKEN(TOK_ENCRYPT, "ENCRYPT", true)
 PARSER_TOKEN(TOK_END, "END", false)
 PARSER_TOKEN(TOK_ENGINE, "ENGINE", true)
 PARSER_TOKEN(TOK_ENTRY_POINT, "ENTRY_POINT", true)
+PARSER_TOKEN(TOK_ERROR, "ERROR", true)
 PARSER_TOKEN(TOK_ESCAPE, "ESCAPE", false)
 PARSER_TOKEN(TOK_EXCEPTION, "EXCEPTION", true)
 PARSER_TOKEN(TOK_EXCESS, "EXCESS", true)
@@ -292,7 +293,7 @@ PARSER_TOKEN(TOK_LIKE, "LIKE", false)
 PARSER_TOKEN(TOK_LIMBO, "LIMBO", true)
 PARSER_TOKEN(TOK_LINGER, "LINGER", true)
 PARSER_TOKEN(TOK_LIST, "LIST", true)
-PARSER_TOKEN(TOK_LISTAGG, "LISTAGG", true)
+PARSER_TOKEN(TOK_LISTAGG, "LISTAGG", false)
 PARSER_TOKEN(TOK_LN, "LN", true)
 PARSER_TOKEN(TOK_LATERAL, "LATERAL", false)
 PARSER_TOKEN(TOK_LOCAL, "LOCAL", false)
@@ -523,6 +524,7 @@ PARSER_TOKEN(TOK_TRIGGER, "TRIGGER", false)
 PARSER_TOKEN(TOK_TRIM, "TRIM", false)
 PARSER_TOKEN(TOK_TRUE, "TRUE", false)
 PARSER_TOKEN(TOK_TRUNC, "TRUNC", true)
+PARSER_TOKEN(TOK_TRUNCATE, "TRUNCATE", false)
 PARSER_TOKEN(TOK_TRUSTED, "TRUSTED", true)
 PARSER_TOKEN(TOK_TWO_PHASE, "TWO_PHASE", true)
 PARSER_TOKEN(TOK_TYPE, "TYPE", true)

src/dsql/AggNodes.cpp

@@ -308,7 +308,7 @@ bool AggNode::dsqlMatch(DsqlCompilerScratch* dsqlScratch, const ExprNode* other,
 	// ASF: We compare name address. That should be ok, as we have only one AggInfo instance
 	// per function.
 	return aggInfo.blr == o->aggInfo.blr && aggInfo.name == o->aggInfo.name &&
-		distinct == o->distinct && dialect1 == o->dialect1 && sort == o->sort;;
+		distinct == o->distinct && dialect1 == o->dialect1;
 }
 
 void AggNode::setParameterName(dsql_par* parameter) const
@@ -448,12 +448,14 @@ bool AggNode::aggPass(thread_db* tdbb, Request* request) const
 			for (auto& nodeOrder : sort->expressions)
 			{
 				dsc toDesc = *(descOrder++);
-				toDesc.dsc_address = data + (IPTR)toDesc.dsc_address;
+				toDesc.dsc_address = data + (IPTR) toDesc.dsc_address;
 				if (const auto fromDsc = EVL_expr(tdbb, request, nodeOrder))
 				{
 					if (IS_INTL_DATA(fromDsc))
+					{
 						INTL_string_to_key(tdbb, INTL_TEXT_TO_INDEX(fromDsc->getTextType()),
 							fromDsc, &toDesc, INTL_KEY_UNIQUE);
+					}
 					else
 						MOV_move(tdbb, fromDsc, &toDesc);
 				}
@@ -465,7 +467,7 @@ bool AggNode::aggPass(thread_db* tdbb, Request* request) const
 			}
 
 			dsc toDesc = asb->desc;
-			toDesc.dsc_address = data + (IPTR)toDesc.dsc_address;
+			toDesc.dsc_address = data + (IPTR) toDesc.dsc_address;
 			MOV_move(tdbb, desc, &toDesc);
 
 			return true;
@@ -522,7 +524,7 @@ dsc* AggNode::execute(thread_db* tdbb, Request* request) const
 			if (distinct)
 				desc.dsc_address = data + (asb->intl ? asb->keyItems[1].getSkdOffset() : 0);
 			else
-				desc.dsc_address = data + (IPTR)asb->desc.dsc_address;
+				desc.dsc_address = data + (IPTR) asb->desc.dsc_address;
 
 			aggPass(tdbb, request, &desc);
 		}
@@ -937,6 +939,20 @@ DmlNode* ListAggNode::parse(thread_db* tdbb, MemoryPool& pool, CompilerScratch*
 	return node;
 }
 
+bool ListAggNode::dsqlMatch(DsqlCompilerScratch* dsqlScratch, const ExprNode* other, bool ignoreMapCast) const
+{
+	if (!AggNode::dsqlMatch(dsqlScratch, other, ignoreMapCast))
+		return false;
+
+	const ListAggNode* o = nodeAs<ListAggNode>(other);
+	fb_assert(o);
+
+	if (dsqlOrderClause || o->dsqlOrderClause)
+		return PASS1_node_match(dsqlScratch, dsqlOrderClause, o->dsqlOrderClause, ignoreMapCast);
+
+	return true;
+}
+
 void ListAggNode::make(DsqlCompilerScratch* dsqlScratch, dsc* desc)
 {
 	DsqlDescMaker::fromNode(dsqlScratch, desc, arg);
@@ -950,20 +966,6 @@ void ListAggNode::genBlr(DsqlCompilerScratch* dsqlScratch)
 	GEN_sort(dsqlScratch, blr_sort, dsqlOrderClause);
 }
 
-AggNode* ListAggNode::pass1(thread_db* tdbb, CompilerScratch* csb)
-{
-	if (sort && distinct)
-	{
-		ValueExprNode* const sortNode = *sort->expressions.begin();
-		if (!arg->sameAs(sortNode, false) || sort->expressions.getCount() > 1)
-		{
-			ERR_post(Arg::Gds(isc_sqlerr) << Arg::Num(-104) << Arg::Gds(isc_dsql_command_err)
-										  << Arg::Gds(isc_distinct_order_by_err));
-		}
-	}
-	return AggNode::pass1(tdbb, csb);
-}
-
 bool ListAggNode::setParameterType(DsqlCompilerScratch* dsqlScratch,
 	std::function<void (dsc*)> makeDesc, bool forceVarChar)
 {

src/dsql/AggNodes.h

@@ -111,10 +111,11 @@ class ListAggNode final : public AggNode
 		holder.add(delimiter);
 	}
 
+	bool dsqlMatch(DsqlCompilerScratch* dsqlScratch, const ExprNode* other, bool ignoreMapCast) const override;
+
 	Firebird::string internalPrint(NodePrinter& printer) const override;
 	void make(DsqlCompilerScratch* dsqlScratch, dsc* desc) override;
 	void genBlr(DsqlCompilerScratch* dsqlScratch) override;
-	AggNode* pass1(thread_db* tdbb, CompilerScratch* csb) override;
 
 	bool setParameterType(DsqlCompilerScratch* dsqlScratch,
 		std::function<void (dsc*)> makeDesc, bool forceVarChar) override;

src/dsql/parse.y

@@ -4723,6 +4723,8 @@ keyword_or_column
 	| GREATEST
 	| LEAST
 	| WITHIN
+	| LISTAGG
+	| TRUNCATE
 	;
 
 col_opt
@@ -10044,7 +10046,7 @@ non_reserved_word
 	| SEARCH_PATH
 	| SCHEMA
 	| UNLIST
-	| LISTAGG
+	| ERROR
 	;
 
 %%

src/include/firebird/impl/msg/jrd.h

@@ -998,4 +998,3 @@ FB_IMPL_MSG(JRD, 995, missing_value_for_format_pattern, -901, "HY", "000", "Cann
 FB_IMPL_MSG(JRD, 996, invalid_name, -901, "HY", "000", "Invalid name: @1")
 FB_IMPL_MSG(JRD, 997, invalid_unqualified_name_list, -901, "HY", "000", "Invalid list of unqualified names: @1")
 FB_IMPL_MSG(JRD, 998, no_user_att_while_restore, -901, "HY", "000", "User attachments are not allowed for the database being restored")
-FB_IMPL_MSG(JRD, 999, distinct_order_by_err, -208, "42", "000", "Sort-key of the ORDER BY specification must match the argument list")

src/include/gen/Firebird.pas

@@ -5844,7 +5844,6 @@ 	IProfilerStatsImpl = class(IProfilerStats)
 	 isc_invalid_name = 335545316;
 	 isc_invalid_unqualified_name_list = 335545317;
 	 isc_no_user_att_while_restore = 335545318;
-	 isc_distinct_order_by_err = 335545319;
 	 isc_gfix_db_name = 335740929;
 	 isc_gfix_invalid_sw = 335740930;
 	 isc_gfix_incmp_sw = 335740932;

src/jrd/optimizer/Optimizer.cpp

@@ -1378,13 +1378,22 @@ void Optimizer::generateAggregateDistincts(MapNode* map)
 			sort_key_def* sort_key = asb->keyItems.getBuffer(asb->intl ? 2 : 1);
 			sort_key->setSkdOffset();
 
+			UCHAR direction = SKD_ascending;
+			if (aggNode->sort)
+			{
+				ValueExprNode* const node = aggNode->sort->expressions.front();
+				const SortDirection sortDir = aggNode->sort->direction.front();
+				if (aggNode->arg->sameAs(node, false) && sortDir == ORDER_DESC)
+					direction = SKD_descending;
+			}
+
 			if (asb->intl)
 			{
 				const USHORT key_length = ROUNDUP(INTL_key_length(tdbb,
 					INTL_TEXT_TO_INDEX(desc->getTextType()), desc->getStringLength()), sizeof(SINT64));
 
 				sort_key->setSkdLength(SKD_bytes, key_length);
-				sort_key->skd_flags = SKD_ascending;
+				sort_key->skd_flags = direction;
 				sort_key->skd_vary_offset = 0;
 
 				++sort_key;
@@ -1412,7 +1421,7 @@ void Optimizer::generateAggregateDistincts(MapNode* map)
 			// 			see AggNode::aggPass() for details; the length remains rounded properly
 			asb->length += sizeof(ULONG);
 
-			sort_key->skd_flags = SKD_ascending;
+			sort_key->skd_flags = direction;
 			asb->impure = csb->allocImpure<impure_agg_sort>();
 			asb->desc = *desc;
 
@@ -1437,8 +1446,8 @@ void Optimizer::generateAggregateSort(AggNode* aggNode)
 	const auto keyCount = aggNode->sort->expressions.getCount() * 2;
 	sort_key_def* sortKey = asb->keyItems.getBuffer(keyCount);
 
-	auto const* direction = aggNode->sort->direction.begin();
-	auto const* nullOrder = aggNode->sort->nullOrder.begin();
+	const auto* direction = aggNode->sort->direction.begin();
+	const auto* nullOrder = aggNode->sort->nullOrder.begin();
 
 	for (auto& node : aggNode->sort->expressions)
 	{