@@ -72,6 +72,7 @@ TEST_SRCS = $(TEST_DIR)/test_main.cpp \
             $(TEST_DIR)/test_mysql_special_expressions.cpp \
             $(TEST_DIR)/test_mysql_cast_types.cpp \
             $(TEST_DIR)/test_mysql_conversion_expressions.cpp \
+            $(TEST_DIR)/test_mysql_insert_sources.cpp \
             $(TEST_DIR)/test_review_ast.cpp \
             $(TEST_DIR)/test_review_transactions.cpp \
             $(TEST_DIR)/test_review_runtime.cpp \
 
@@ -16,9 +16,11 @@ MySQL query expressions also preserve explicit `TABLE name` and `VALUES ROW(...)
 
 Top-level MySQL `WITH` and `WITH RECURSIVE` also retain structured UPDATE and DELETE main statements, including join/comma UPDATEs, both multi-table DELETE target-list forms, quoted CTE names, partitions, statement options, and single-table ordering/limits. CTE bodies and subqueries remain query-only; DML inside them and a leading `WITH ... INSERT` are rejected. UPDATE assignments require one-, two-, or three-part column names and accept both `=` and `:=` (emitted as `=`), including `DEFAULT` values. UPDATE options follow `LOW_PRIORITY` then `IGNORE`; DELETE options may appear in any order and repeat. The [8.4.8 UPDATE/DELETE productions](https://github.com/mysql/mysql-server/blob/0896fcd61dec11a0904166911a0126f59daaa1bf/sql/sql_yacc.yy#L13389) and [9.7.2 UPDATE/DELETE productions](https://github.com/mysql/mysql-server/blob/008e09c2834b98143a8c067d4d225c90953050cf/sql/sql_yacc.yy#L13945) attach an optional WITH clause to each main statement. Routing metadata identifies the main DML target, including wildcard DELETE targets. JOIN conditions require complete ON/USING operands and retain quoted USING column names. These are parser/AST guarantees; the local engine rejects CTE-prefixed DML.
 
+MySQL INSERT constructors and SET sources retain `AS row_alias` and optional column aliases before `ON DUPLICATE KEY UPDATE`, including quoted names and doubled-backtick escapes. Alias emission separates the row name from its column-list parentheses. With `IGNORE_SPACE` off, native function names such as `COUNT` and `MAX` require this separator when used as unquoted row aliases; adjacent parentheses turn them into function tokens and are rejected. Explicit alias-column parentheses require a nonempty identifier list; the native empty production means that the entire list is omitted. `VALUE` emits canonically as `VALUES`, and `:=` assignment operators emit as `=`. SET and duplicate-key assignments require complete column targets and values, including qualified targets and `DEFAULT`. INSERT also accepts partition lists and query sources containing set operations, parentheses, WITH, TABLE, and `VALUES ROW(...)`. Row aliases belong to INSERT's VALUES/SET source productions; REPLACE and query sources do not accept this suffix. REPLACE also excludes duplicate-key updates. The [8.4.8 INSERT/REPLACE productions](https://github.com/mysql/mysql-server/blob/0896fcd61dec11a0904166911a0126f59daaa1bf/sql/sql_yacc.yy#L13064) and [9.7.2 INSERT/REPLACE productions](https://github.com/mysql/mysql-server/blob/008e09c2834b98143a8c067d4d225c90953050cf/sql/sql_yacc.yy#L13620), with the corpus's constructor, alias-list, and assignment anchors, define these bounded parser/AST guarantees. Parameterization remains conservative for row aliases. Local INSERT execution and direct INSERT planning reject unsupported row aliases, query sources, and partition selection. Native checks use EXPLAIN; they do not establish local-engine execution support.
+
 The native grammar permits empty `VALUES ROW()`, `VALUES ROW(DEFAULT)`, and repeated locking clauses. Both pinned servers return semantic errors 3942, 3943, and 3569 respectively for the curated standalone or overlapping-lock witnesses. These remain `SERVER_ERROR` observations, not syntax rejections, while ParserSQL preserves the grammar-valid ASTs. `TABLE` takes a table identifier and query-level tails; aliases, partitions, index hints, and WHERE clauses directly after its table name are malformed controls. The [8.4.8 query-primary productions](https://github.com/mysql/mysql-server/blob/0896fcd61dec11a0904166911a0126f59daaa1bf/sql/sql_yacc.yy#L9831) and [9.7.2 query-primary productions](https://github.com/mysql/mysql-server/blob/008e09c2834b98143a8c067d4d225c90953050cf/sql/sql_yacc.yy#L10145) include both constructors. The corpus records exact anchors for their rows, subquery contexts, locking, CTEs, and malformed controls.
 
-Table identifiers and aliases use the common 259-word reserved baseline observed in the pinned 8.4.8 and 9.7.2 releases. ParserSQL has no MySQL version setting, so the five 9.7-only reserved additions (`CUBE`, `EXTERNAL`, `LIBRARY`, `QUALIFY`, and `TABLESAMPLE`) remain accepted as names. Strict version-specific identifier rules remain a gap.
+Table identifiers and aliases use the common 259-word reserved baseline observed in the pinned 8.4.8 and 9.7.2 releases. ParserSQL has no MySQL version setting, so the five 9.7-only reserved additions (`CUBE`, `EXTERNAL`, `LIBRARY`, `QUALIFY`, and `TABLESAMPLE`) remain accepted as names. Strict version-specific identifier rules remain a gap. The INSERT alias function-token check likewise uses the 35 names shared by both pins; the additional 9.7 `JSON_DUALITY_OBJECT` function-token distinction remains outside that common baseline.
 
 Numeric-looking unquoted qualified table names such as `db.123` are valid natively but remain unsupported: the tokenizer reads `.123` as a numeric literal. ParserSQL rejects these forms; earlier behavior could emit an incorrect alias. Lossless parsing of this identifier edge remains a gap.
 
@@ -54,7 +56,7 @@ Malformed controls are handled separately: a complete AST is `MALFORMED_ACCEPTED
 
 Every fixture freezes one exact expected outcome. Known gaps do not make arbitrary failures acceptable. Both regression and unexpected improvement cause exit status 1, so a fixed gap must be reviewed and explicitly promoted in the corpus. When promoting, add an independently derived `expected_sql`, set `known_gap=false`, and require `CHECKED`. Harness/transport/source errors return 2. Exit status 0 means the expected baseline matched, including its documented known gaps.
 
-The 212-case baseline has 110 `CHECKED` witnesses and 92 `REJECTED` negative controls. Its 10 remaining known gaps are five `TRAILING_INPUT`, three `NO_AST`, one `PARTIAL`, and one `EMISSION_MISMATCH`. CTE-prefixed UPDATE/DELETE, EXTRACT, substring keyword forms, MATCH, casts/conversions, charset-introduced literals, and COLLATE now pass complete parse, exact emission, and stable reparse checks. Source verification checks 212 exact source anchors per pinned release. These counts describe this corpus only.
+The 260-case baseline has 127 `CHECKED` witnesses and 124 `REJECTED` negative controls. Its nine remaining known gaps are four `TRAILING_INPUT`, three `NO_AST`, one `PARTIAL`, and one `EMISSION_MISMATCH`. CTE-prefixed UPDATE/DELETE, EXTRACT, substring keyword forms, MATCH, casts/conversions, charset-introduced literals, COLLATE, and the curated INSERT/REPLACE forms now pass complete parse, exact emission, and stable reparse checks. Source verification checks 268 exact source anchors per pinned release. These counts describe this corpus only.
 
 The JSON report includes all original and reparsed records, source verification counts, pinned revisions, SQL-mode metadata, and SHA-256 hashes of the corpus and probe. Counts apply only to these witnesses. The probe uses hex-line input and hex fields in its TSV response, preserving tabs, newlines and embedded NUL bytes. It emits while both the input string and parser arena are alive; Python reparses owned copies in a second probe invocation.
 
@@ -73,9 +75,9 @@ Each curated case explicitly chooses direct query execution, `EXPLAIN` for DML (
 
 Native results are stored separately in `native_oracle`: `ACCEPTED`, `SYNTAX_REJECTED` (1064/1149), `SERVER_ERROR`, `CLIENT_ERROR`, or **NOT VERIFIED**. For example, a derived table missing its required alias returns error 1248; this remains a server error, even though the corpus requires ParserSQL to reject that input. Accepted invalid controls and syntax-rejected valid controls count as expectation mismatches, using per-version validity when specified. Server errors remain explicitly unresolved; client errors return status 2. Thus exit status 0 still does not mean all native cases were verified. The default corpus metadata remains **NOT VERIFIED**, while the optional report describes the actual native observations.
 
-A native check of all 212 witnesses under `ONLY_FULL_GROUP_BY,STRICT_TRANS_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ERROR_FOR_DIVISION_BY_ZERO,NO_ENGINE_SUBSTITUTION` found no native expectation mismatches. MySQL 8.4.8 accepted 110, syntax-rejected 94, and returned five server errors (two 1248 missing-alias errors, 3942, 3943, and 3569); MySQL 9.7.2 accepted 112, syntax-rejected 90, and returned seven server errors (the same five plus 1235 and 3889). Both runs skipped the three DDL/procedure witnesses and cleaned up their owned schemas. All shared positive query/DML witnesses except the three explicitly grammar-valid semantic-error controls were accepted, including partition selection, index hints, query constructors, recursive CTEs, CTE-prefixed UPDATE/DELETE, all twenty EXTRACT units, substring aliases, MATCH modes, casts/conversions, charset-introduced literals, and COLLATE operands. Re-run the command on each isolated socket to reproduce current observations; reports are intentionally not committed.
+A native check of all 260 witnesses under `ONLY_FULL_GROUP_BY,STRICT_TRANS_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ERROR_FOR_DIVISION_BY_ZERO,NO_ENGINE_SUBSTITUTION` found no native expectation mismatches. MySQL 8.4.8 accepted 126, syntax-rejected 126, and returned five server errors (two 1248 missing-alias errors, 3942, 3943, and 3569); MySQL 9.7.2 accepted 128, syntax-rejected 122, and returned seven server errors (the same five plus 1235 and 3889). Both runs skipped the three DDL/procedure witnesses and cleaned up their owned schemas. All shared positive query/DML witnesses except the three explicitly grammar-valid semantic-error controls were accepted, including partition selection, index hints, query constructors, recursive CTEs, CTE-prefixed UPDATE/DELETE, all twenty EXTRACT units, substring aliases, MATCH modes, casts/conversions, charset-introduced literals, COLLATE operands, and INSERT constructors, aliases, assignments, and query sources. The 48 added INSERT/REPLACE witnesses contributed 16 accepted EXPLAIN statements and 32 syntax errors in each version. Re-run the command on each isolated socket to reproduce current observations; reports are intentionally not committed.
 
-The corpus is deliberately selected and is not a representative workload. It contains 208 shared witnesses and four version-specific query witnesses; this does not imply identical dialects in 8.4.8 and 9.7.2. The version-specific entries record `applicability`, source productions for both versions, `native_validity` per version, and observed native outcomes. Their generic `validity=valid` denotes the 9.7.2 grammar target. The native helper selects the intended validity using the actual server version, so an expected 8.4.8 syntax rejection is not treated as a corpus error. Stored native observations are context, not substitutes for a current run. These additions remain ParserSQL gaps. The following differences include those four query witnesses and two separate DDL spot checks:
+The corpus is deliberately selected and is not a representative workload. It contains 256 shared witnesses and four version-specific query witnesses; this does not imply identical dialects in 8.4.8 and 9.7.2. The version-specific entries record `applicability`, source productions for both versions, `native_validity` per version, and observed native outcomes. Their generic `validity=valid` denotes the 9.7.2 grammar target. The native helper selects the intended validity using the actual server version, so an expected 8.4.8 syntax rejection is not treated as a corpus error. Stored native observations are context, not substitutes for a current run. These additions remain ParserSQL gaps. The following differences include those four query witnesses and two separate DDL spot checks:
 
 | Reduced feature | MySQL 8.4.8 observation | MySQL 9.7.2 observation |
 | --- | --- | --- |
 
@@ -206,6 +206,10 @@ class DmlPlanBuilder {
         if (!ast) return false;
         if constexpr (D == sql_parser::Dialect::PostgreSQL) {
             if (!PlanBuilder<D>::supports_dml_features(ast) || unsupported_dml(ast)) return false;
+        } else if (ast->type == sql_parser::NodeType::NODE_INSERT_STMT) {
+            // INSERT plans must not discard newly parsed aliases or sources.
+            // UPDATE/DELETE retain their existing original-AST routing paths.
+            if (!PlanBuilder<D>::supports_dml_features(ast) || unsupported_dml(ast)) return false;
         }
         return true;
     }
 
@@ -194,6 +194,7 @@ class PlanBuilder {
             case NodeType::NODE_MYSQL_CHARSET_LITERAL:
             case NodeType::NODE_MYSQL_COLLATE:
             case NodeType::NODE_MYSQL_CONVERT:
+            case NodeType::NODE_MYSQL_INSERT_ALIAS:
             case NodeType::NODE_MYSQL_EXTRACT:
             case NodeType::NODE_MYSQL_SUBSTRING:
             case NodeType::NODE_MYSQL_MATCH:
 
@@ -361,6 +361,7 @@ enum class NodeType : uint16_t {
     NODE_MYSQL_CHARSET_LITERAL, // value=charset introducer; string/hex/bit literal children
     NODE_MYSQL_COLLATE, // value=collation source spelling; one expression child
     NODE_MYSQL_CONVERT, // expression, type/charset metadata; flag 1=USING form
+    NODE_MYSQL_INSERT_ALIAS, // row identifier, optional nonempty column identifier list
 
 };
 
 
@@ -98,6 +98,21 @@ class Emitter {
             case NodeType::NODE_VALUES_CLAUSE:   emit_values_clause(node); break;
             case NodeType::NODE_VALUES_ROW:      emit_values_row(node); break;
             case NodeType::NODE_INSERT_SET_CLAUSE: emit_insert_set_clause(node); break;
+            case NodeType::NODE_MYSQL_INSERT_ALIAS: {
+                sb_.append("AS ");
+                const auto* name = node->first_child;
+                if (name) emit_node(name);
+                if (name && name->next_sibling) {
+                    // COUNT (a) is an alias; COUNT(a) is a native function token.
+                    sb_.append(" (");
+                    for (auto* column = name->next_sibling; column; column = column->next_sibling) {
+                        if (column != name->next_sibling) sb_.append(", ");
+                        emit_node(column);
+                    }
+                    sb_.append_char(')');
+                }
+                break;
+            }
             case NodeType::NODE_ON_DUPLICATE_KEY: emit_on_duplicate_key(node); break;
             case NodeType::NODE_ON_CONFLICT:     emit_on_conflict(node); break;
             case NodeType::NODE_CONFLICT_TARGET: emit_conflict_target(node); break;
Original file line number	Diff line number	Diff line change
`@@ -206,6 +206,10 @@ class DmlPlanBuilder {`
`206`	`206`	`if (!ast) return false;`
`207`	`207`	`if constexpr (D == sql_parser::Dialect::PostgreSQL) {`
`208`	`208`	`if (!PlanBuilder<D>::supports_dml_features(ast) \|\| unsupported_dml(ast)) return false;`
	`209`	`+ } else if (ast->type == sql_parser::NodeType::NODE_INSERT_STMT) {`
	`210`	`+ // INSERT plans must not discard newly parsed aliases or sources.`
	`211`	`+ // UPDATE/DELETE retain their existing original-AST routing paths.`
	`212`	`+ if (!PlanBuilder<D>::supports_dml_features(ast) \|\| unsupported_dml(ast)) return false;`
`209`	`213`	`}`
`210`	`214`	`return true;`
`211`	`215`	`}`