Add the support of OPTIMIZE expression

parser/ast.go

@@ -3049,6 +3049,78 @@ func (f *FormatExpr) String(level int) string {
 	return "FORMAT " + f.Format.String(level)
 }
 
+type OptimizeExpr struct {
+	OptimizePos  Pos
+	StatementEnd Pos
+	Table        *TableIdentifier
+	OnCluster    *OnClusterExpr
+	Partition    *PartitionExpr
+	HasFinal     bool
+	Deduplicate  *DeduplicateExpr
+}
+
+func (o *OptimizeExpr) Pos() Pos {
+	return o.OptimizePos
+}
+
+func (o *OptimizeExpr) End() Pos {
+	return o.StatementEnd
+}
+
+func (o *OptimizeExpr) String(level int) string {
+	var builder strings.Builder
+	builder.WriteString("OPTIMIZE TABLE ")
+	builder.WriteString(o.Table.String(level))
+	if o.OnCluster != nil {
+		builder.WriteString(NewLine(level))
+		builder.WriteString(o.OnCluster.String(level))
+	}
+	if o.Partition != nil {
+		builder.WriteString(NewLine(level))
+		builder.WriteString(o.Partition.String(level))
+	}
+	if o.HasFinal {
+		builder.WriteString(" FINAL")
+	}
+	if o.Deduplicate != nil {
+		builder.WriteString(o.Deduplicate.String(level))
+	}
+	return builder.String()
+}
+
+type DeduplicateExpr struct {
+	DeduplicatePos Pos
+	By             *ColumnExprList
+	Except         *ColumnExprList
+}
+
+func (d *DeduplicateExpr) Pos() Pos {
+	return d.DeduplicatePos
+}
+
+func (d *DeduplicateExpr) End() Pos {
+	if d.By != nil {
+		return d.By.End()
+	} else if d.Except != nil {
+		return d.Except.End()
+	}
+	return d.DeduplicatePos + Pos(len(KeywordDeduplicate))
+}
+
+func (d *DeduplicateExpr) String(level int) string {
+	var builder strings.Builder
+	builder.WriteString(" DEDUPLICATE")
+	if d.By != nil {
+		builder.WriteString(" BY ")
+		builder.WriteString(d.By.String(level))
+	}
+	if d.Except != nil {
+		builder.WriteString(" EXCEPT ")
+		builder.WriteString(d.Except.String(level))
+	}
+	return builder.String()
+}
+
 type SystemExpr struct {
 	SystemPos Pos
 	Expr      Expr

parser/keyword.go

@@ -60,6 +60,7 @@ const (
 	KeywordEnd          = "END"
 	KeywordEngine       = "ENGINE"
 	KeywordEvents       = "EVENTS"
+	KeywordExcept       = "EXCEPT"
 	KeywordExists       = "EXISTS"
 	KeywordExplain      = "EXPLAIN"
 	KeywordExpression   = "EXPRESSION"
@@ -261,6 +262,7 @@ var keywords = NewSet(
 	KeywordEnd,
 	KeywordEngine,
 	KeywordEvents,
+	KeywordExcept,
 	KeywordExists,
 	KeywordExplain,
 	KeywordExpression,

parser/parse_system.go

@@ -198,6 +198,90 @@ func (p *Parser) parseSystemDropExpr(pos Pos) (*SystemDropExpr, error) {
 	}
 }
 
+func (p *Parser) tryParseDeduplicateExpr(pos Pos) (*DeduplicateExpr, error) {
+	if !p.matchKeyword(KeywordDeduplicate) {
+		return nil, nil
+	}
+	return p.parseDeduplicateExpr(pos)
+}
+
+func (p *Parser) parseDeduplicateExpr(pos Pos) (*DeduplicateExpr, error) {
+	if err := p.consumeKeyword(KeywordDeduplicate); err != nil {
+		return nil, err
+	}
+	if p.tryConsumeKeyword(KeywordBy) == nil {
+		return &DeduplicateExpr{
+			DeduplicatePos: pos,
+		}, nil
+	}
+
+	by, err := p.parseColumnExprList(p.Pos())
+	if err != nil {
+		return nil, err
+	}
+	var except *ColumnExprList
+	if p.tryConsumeKeyword(KeywordExcept) != nil {
+		except, err = p.parseColumnExprList(p.Pos())
+		if err != nil {
+			return nil, err
+		}
+	}
+	return &DeduplicateExpr{
+		DeduplicatePos: pos,
+		By:             by,
+		Except:         except,
+	}, nil
+}
+
+func (p *Parser) parseOptimizeExpr(pos Pos) (*OptimizeExpr, error) {
+	if err := p.consumeKeyword(KeywordOptimize); err != nil {
+		return nil, err
+	}
+	if err := p.consumeKeyword(KeywordTable); err != nil {
+		return nil, err
+	}
+	table, err := p.parseTableIdentifier(p.Pos())
+	if err != nil {
+		return nil, err
+	}
+	statmentEnd := table.End()
+	onCluster, err := p.tryParseOnCluster(p.Pos())
+	if err != nil {
+		return nil, err
+	}
+	if onCluster != nil {
+		statmentEnd = onCluster.End()
+	}
+	partitionExpr, err := p.tryParsePartitionExpr(p.Pos())
+	if err != nil {
+		return nil, err
+	}
+	if partitionExpr != nil {
+		statmentEnd = partitionExpr.End()
+	}
+
+	hasFinal := false
+	lastPos := p.Pos()
+	if p.tryConsumeKeyword(KeywordFinal) != nil {
+		hasFinal = true
+		statmentEnd = lastPos
+	}
+	deduplicate, err := p.tryParseDeduplicateExpr(p.Pos())
+	if err != nil {
+		return nil, err
+	}
+
+	return &OptimizeExpr{
+		OptimizePos:  pos,
+		StatementEnd: statmentEnd,
+		Table:        table,
+		OnCluster:    onCluster,
+		Partition:    partitionExpr,
+		HasFinal:     hasFinal,
+		Deduplicate:  deduplicate,
+	}, nil
+}
+
 func (p *Parser) parseSystemExpr(pos Pos) (*SystemExpr, error) {
 	if err := p.consumeKeyword(KeywordSystem); err != nil {
 		return nil, err

parser/parser_table.go

@@ -904,6 +904,8 @@ func (p *Parser) parseStatement(pos Pos) (Expr, error) {
 		expr, err = p.parseSetExpr(pos)
 	case p.matchKeyword(KeywordSystem):
 		expr, err = p.parseSystemExpr(pos)
+	case p.matchKeyword(KeywordOptimize):
+		expr, err = p.parseOptimizeExpr(pos)
 	default:
 		return nil, fmt.Errorf("unexpected token: %q", p.last().String)
 	}

parser/testdata/ddl/format/optimize.sql

@@ -0,0 +1,19 @@
+-- Origin SQL:
+OPTIMIZE TABLE table DEDUPLICATE; -- all columns
+OPTIMIZE TABLE table DEDUPLICATE BY *; -- excludes MATERIALIZED and ALIAS columns
+OPTIMIZE TABLE table DEDUPLICATE BY colX,colY,colZ;
+OPTIMIZE TABLE table DEDUPLICATE BY * EXCEPT colX;
+OPTIMIZE TABLE table DEDUPLICATE BY * EXCEPT (colX, colY);
+OPTIMIZE TABLE table DEDUPLICATE BY COLUMNS('column-matched-by-regex');
+OPTIMIZE TABLE table DEDUPLICATE BY COLUMNS('column-matched-by-regex') EXCEPT colX;
+OPTIMIZE TABLE table DEDUPLICATE BY COLUMNS('column-matched-by-regex') EXCEPT (colX, colY);
+
+-- Format SQL:
+OPTIMIZE TABLE table DEDUPLICATE;
+OPTIMIZE TABLE table DEDUPLICATE BY *;
+OPTIMIZE TABLE table DEDUPLICATE BY colX, colY, colZ;
+OPTIMIZE TABLE table DEDUPLICATE BY * EXCEPT colX;
+OPTIMIZE TABLE table DEDUPLICATE BY * EXCEPT (colX, colY);
+OPTIMIZE TABLE table DEDUPLICATE BY COLUMNS('column-matched-by-regex');
+OPTIMIZE TABLE table DEDUPLICATE BY COLUMNS('column-matched-by-regex') EXCEPT colX;
+OPTIMIZE TABLE table DEDUPLICATE BY COLUMNS('column-matched-by-regex') EXCEPT (colX, colY);

parser/testdata/ddl/optimize.sql

@@ -0,0 +1,8 @@
+OPTIMIZE TABLE table DEDUPLICATE; -- all columns
+OPTIMIZE TABLE table DEDUPLICATE BY *; -- excludes MATERIALIZED and ALIAS columns
+OPTIMIZE TABLE table DEDUPLICATE BY colX,colY,colZ;
+OPTIMIZE TABLE table DEDUPLICATE BY * EXCEPT colX;
+OPTIMIZE TABLE table DEDUPLICATE BY * EXCEPT (colX, colY);
+OPTIMIZE TABLE table DEDUPLICATE BY COLUMNS('column-matched-by-regex');
+OPTIMIZE TABLE table DEDUPLICATE BY COLUMNS('column-matched-by-regex') EXCEPT colX;
+OPTIMIZE TABLE table DEDUPLICATE BY COLUMNS('column-matched-by-regex') EXCEPT (colX, colY);