lateral join, unnest, parquet encoding config support

Author: emrberk

src/autocomplete/suggestion-builder.ts

@@ -80,6 +80,9 @@ function getAllColumns(schema: SchemaInfo): ColumnWithTable[] {
  * into compound suggestions (e.g., "Left" → "LEFT JOIN") instead of
  * suggesting bare "LEFT" which is incomplete on its own.
  */
+// Keywords that are valid but rarely the primary intent — suggest at MediumLow.
+const MEDIUM_LOW_KEYWORDS = new Set(["Unnest"])
+
 const JOIN_COMPOUND_MAP = new Map<string, string>([
   ["Left", "LEFT JOIN"],
   ["Inner", "INNER JOIN"],
@@ -170,12 +173,17 @@ export function buildSuggestions(
     // Functions are suggested separately in the functions loop below.
     const kind = SuggestionKind.Keyword
 
+    // Some keywords are valid but rarely typed directly — suggest them lower.
+    const priority = MEDIUM_LOW_KEYWORDS.has(name)
+      ? SuggestionPriority.MediumLow
+      : SuggestionPriority.Medium
+
     suggestions.push({
       label: keyword,
       kind,
       insertText: keyword,
       filterText: keyword.toLowerCase(),
-      priority: SuggestionPriority.Medium,
+      priority,
     })
   }
 

src/parser/ast.ts

@@ -233,6 +233,18 @@ export interface MaterializedViewPeriod extends AstNode {
   sampleByInterval?: boolean
 }
 
+export interface ParquetConfig extends AstNode {
+  type: "parquetConfig"
+  /** Encoding: PLAIN, RLE_DICTIONARY, DELTA_BINARY_PACKED, DELTA_LENGTH_BYTE_ARRAY, DEFAULT */
+  encoding?: string
+  /** Compression codec: UNCOMPRESSED, SNAPPY, GZIP, BROTLI, ZSTD, LZ4_RAW */
+  compression?: string
+  /** Compression level, e.g. 3 for ZSTD(3) */
+  compressionLevel?: number
+  /** Whether BLOOM_FILTER is enabled */
+  bloomFilter?: boolean
+}
+
 export interface ColumnDefinition extends AstNode {
   type: "columnDefinition"
   name: string
@@ -245,6 +257,8 @@ export interface ColumnDefinition extends AstNode {
   indexed?: boolean
   /** INDEX CAPACITY value */
   indexCapacity?: number
+  /** PARQUET encoding/compression/bloom filter config */
+  parquetConfig?: ParquetConfig
 }
 
 export interface AlterTableStatement extends AstNode {
@@ -411,9 +425,12 @@ export interface AlterColumnAction {
     | "cache"
     | "nocache"
     | "symbolCapacity"
+    | "setParquet"
   newType?: string
   capacity?: number
   cache?: boolean
+  /** PARQUET config for SET PARQUET(...) */
+  parquetConfig?: ParquetConfig
 }
 
 export interface DropPartitionAction {
@@ -820,10 +837,38 @@ export interface TableFunctionCall extends AstNode {
   args: Expression[]
 }
 
+export interface UnnestArg extends AstNode {
+  type: "unnestArg"
+  expression: Expression
+  /** JSON UNNEST column definitions: COLUMNS(name TYPE, ...) */
+  columns?: UnnestColumnDef[]
+}
+
+export interface UnnestColumnDef extends AstNode {
+  type: "unnestColumnDef"
+  name: string
+  dataType: string
+}
+
+export interface UnnestSource extends AstNode {
+  type: "unnest"
+  args: UnnestArg[]
+  withOrdinality?: boolean
+}
+
 export interface TableRef extends AstNode {
   type: "tableRef"
-  table: QualifiedName | SelectStatement | TableFunctionCall | ShowStatement
+  table:
+    | QualifiedName
+    | SelectStatement
+    | TableFunctionCall
+    | ShowStatement
+    | UnnestSource
+  /** Whether this table ref was preceded by LATERAL in comma-separated FROM */
+  lateral?: boolean
   alias?: string
+  /** Column alias list for UNNEST sources, e.g. UNNEST(arr) u(val, ord) */
+  columnAliases?: string[]
   joins?: JoinClause[]
   timestampDesignation?: string
 }
@@ -840,6 +885,8 @@ export interface JoinClause extends AstNode {
     | "window"
     | "horizon"
   outer?: boolean
+  /** Whether LATERAL was specified after JOIN keyword */
+  lateral?: boolean
   table: TableRef
   on?: Expression
   /** Tolerance interval for ASOF and LT joins (e.g., "1h", "30s") */

src/parser/cst-types.d.ts

@@ -186,9 +186,61 @@ export interface FromClauseCstNode extends CstNode {
 }
 
 export type FromClauseCstChildren = {
-  tableRef: (TableRefCstNode)[];
+  fromSource: (FromSourceCstNode)[];
   joinClause?: JoinClauseCstNode[];
   Comma?: IToken[];
+  Lateral?: IToken[];
+};
+
+export interface FromSourceCstNode extends CstNode {
+  name: "fromSource";
+  children: FromSourceCstChildren;
+}
+
+export type FromSourceCstChildren = {
+  unnestClause?: UnnestClauseCstNode[];
+  tableRef?: TableRefCstNode[];
+};
+
+export interface UnnestArgCstNode extends CstNode {
+  name: "unnestArg";
+  children: UnnestArgCstChildren;
+}
+
+export type UnnestArgCstChildren = {
+  expression: ExpressionCstNode[];
+  Columns?: IToken[];
+  LParen?: IToken[];
+  unnestColumnDef?: (UnnestColumnDefCstNode)[];
+  Comma?: IToken[];
+  RParen?: IToken[];
+};
+
+export interface UnnestColumnDefCstNode extends CstNode {
+  name: "unnestColumnDef";
+  children: UnnestColumnDefCstChildren;
+}
+
+export type UnnestColumnDefCstChildren = {
+  identifier: IdentifierCstNode[];
+  dataType: DataTypeCstNode[];
+};
+
+export interface UnnestClauseCstNode extends CstNode {
+  name: "unnestClause";
+  children: UnnestClauseCstChildren;
+}
+
+export type UnnestClauseCstChildren = {
+  Unnest: IToken[];
+  LParen: (IToken)[];
+  unnestArg: (UnnestArgCstNode)[];
+  Comma?: (IToken)[];
+  RParen: (IToken)[];
+  With?: IToken[];
+  Ordinality?: IToken[];
+  As?: IToken[];
+  identifier?: (IdentifierCstNode)[];
 };
 
 export interface ImplicitSelectBodyCstNode extends CstNode {
@@ -368,6 +420,7 @@ export type StandardJoinCstChildren = {
   Inner?: IToken[];
   Cross?: IToken[];
   Join: IToken[];
+  Lateral?: IToken[];
   tableRef: TableRefCstNode[];
   On?: IToken[];
   expression?: ExpressionCstNode[];
@@ -911,6 +964,50 @@ export type ColumnDefinitionCstChildren = {
   Cache?: IToken[];
   Nocache?: IToken[];
   Index?: IToken[];
+  parquetConfig?: ParquetConfigCstNode[];
+};
+
+export interface ParquetConfigCstNode extends CstNode {
+  name: "parquetConfig";
+  children: ParquetConfigCstChildren;
+}
+
+export type ParquetConfigCstChildren = {
+  Parquet: IToken[];
+  LParen: (IToken)[];
+  BloomFilter?: (IToken)[];
+  parquetEncoding?: ParquetEncodingCstNode[];
+  Comma?: (IToken)[];
+  parquetCompression?: ParquetCompressionCstNode[];
+  NumberLiteral?: IToken[];
+  RParen: (IToken)[];
+};
+
+export interface ParquetEncodingCstNode extends CstNode {
+  name: "parquetEncoding";
+  children: ParquetEncodingCstChildren;
+}
+
+export type ParquetEncodingCstChildren = {
+  Plain?: IToken[];
+  RleDictionary?: IToken[];
+  DeltaBinaryPacked?: IToken[];
+  DeltaLengthByteArray?: IToken[];
+  Default?: IToken[];
+};
+
+export interface ParquetCompressionCstNode extends CstNode {
+  name: "parquetCompression";
+  children: ParquetCompressionCstChildren;
+}
+
+export type ParquetCompressionCstChildren = {
+  Uncompressed?: IToken[];
+  Snappy?: IToken[];
+  Gzip?: IToken[];
+  Brotli?: IToken[];
+  Zstd?: IToken[];
+  Lz4Raw?: IToken[];
 };
 
 export interface CastDefinitionCstNode extends CstNode {
@@ -1140,11 +1237,12 @@ export type AlterTableActionCstChildren = {
   Nocache?: (IToken)[];
   Index?: (IToken)[];
   Symbol?: IToken[];
+  Set?: (IToken)[];
+  parquetConfig?: ParquetConfigCstNode[];
   Attach?: IToken[];
   Detach?: IToken[];
   Squash?: IToken[];
   Partitions?: IToken[];
-  Set?: IToken[];
   Param?: IToken[];
   tableParam?: (TableParamCstNode)[];
   Ttl?: IToken[];
@@ -2453,6 +2551,10 @@ export interface ICstNodeVisitor<IN, OUT> extends ICstVisitor<IN, OUT> {
   selectItem(children: SelectItemCstChildren, param?: IN): OUT;
   qualifiedStar(children: QualifiedStarCstChildren, param?: IN): OUT;
   fromClause(children: FromClauseCstChildren, param?: IN): OUT;
+  fromSource(children: FromSourceCstChildren, param?: IN): OUT;
+  unnestArg(children: UnnestArgCstChildren, param?: IN): OUT;
+  unnestColumnDef(children: UnnestColumnDefCstChildren, param?: IN): OUT;
+  unnestClause(children: UnnestClauseCstChildren, param?: IN): OUT;
   implicitSelectBody(children: ImplicitSelectBodyCstChildren, param?: IN): OUT;
   implicitSelectStatement(children: ImplicitSelectStatementCstChildren, param?: IN): OUT;
   tableRef(children: TableRefCstChildren, param?: IN): OUT;
@@ -2498,6 +2600,9 @@ export interface ICstNodeVisitor<IN, OUT> extends ICstVisitor<IN, OUT> {
   materializedViewPeriod(children: MaterializedViewPeriodCstChildren, param?: IN): OUT;
   materializedViewPartition(children: MaterializedViewPartitionCstChildren, param?: IN): OUT;
   columnDefinition(children: ColumnDefinitionCstChildren, param?: IN): OUT;
+  parquetConfig(children: ParquetConfigCstChildren, param?: IN): OUT;
+  parquetEncoding(children: ParquetEncodingCstChildren, param?: IN): OUT;
+  parquetCompression(children: ParquetCompressionCstChildren, param?: IN): OUT;
   castDefinition(children: CastDefinitionCstChildren, param?: IN): OUT;
   indexDefinition(children: IndexDefinitionCstChildren, param?: IN): OUT;
   tableParamName(children: TableParamNameCstChildren, param?: IN): OUT;

src/parser/lexer.ts

@@ -27,6 +27,7 @@ import {
   Between,
   Binary,
   Boolean,
+  BloomFilter,
   By,
   Bypass,
   Byte,
@@ -63,6 +64,8 @@ import {
   Dedup,
   Default,
   DefaultTransactionReadOnly,
+  DeltaBinaryPacked,
+  DeltaLengthByteArray,
   Deferred,
   Delay,
   Delete,
@@ -128,6 +131,7 @@ import {
   Keep,
   Key,
   Keys,
+  Lateral,
   Latest,
   Left,
   Length,
@@ -172,6 +176,7 @@ import {
   Option,
   Or,
   Order,
+  Ordinality,
   Others,
   Outer,
   Over,
@@ -188,6 +193,7 @@ import {
   Period,
   Permissions,
   Pivot,
+  Plain,
   Prepare,
   Preceding,
   Prev,
@@ -209,6 +215,7 @@ import {
   Resume,
   Revoke,
   Right,
+  RleDictionary,
   Row,
   RowGroupSize,
   Rows,
@@ -257,6 +264,7 @@ import {
   Unbounded,
   Union,
   Unlock,
+  Unnest,
   Unpivot,
   Update,
   Upsert,
@@ -317,6 +325,7 @@ export {
   Between,
   Binary,
   Boolean,
+  BloomFilter,
   By,
   Bypass,
   Byte,
@@ -353,6 +362,8 @@ export {
   Dedup,
   Default,
   DefaultTransactionReadOnly,
+  DeltaBinaryPacked,
+  DeltaLengthByteArray,
   Deferred,
   Delay,
   Delete,
@@ -418,6 +429,7 @@ export {
   Keep,
   Key,
   Keys,
+  Lateral,
   Latest,
   Left,
   Length,
@@ -462,6 +474,7 @@ export {
   Option,
   Or,
   Order,
+  Ordinality,
   Others,
   Outer,
   Over,
@@ -478,6 +491,7 @@ export {
   Period,
   Permissions,
   Pivot,
+  Plain,
   Prepare,
   Preceding,
   Prev,
@@ -499,6 +513,7 @@ export {
   Resume,
   Revoke,
   Right,
+  RleDictionary,
   Row,
   RowGroupSize,
   Rows,
@@ -547,6 +562,7 @@ export {
   Unbounded,
   Union,
   Unlock,
+  Unnest,
   Unpivot,
   Update,
   Upsert,