Support additional DuckDB integer types such as HUGEINT, UHUGEINT, etc (apache#1797)

Co-authored-by: Alexander Beedie <[email protected]>

Author: alexander-beedie

src/ast/data_type.rs

@@ -1276,9 +1276,9 @@ impl fmt::Display for IndexOption {
     }
 }
 
-/// [Postgres] unique index nulls handling option: `[ NULLS [ NOT ] DISTINCT ]`
+/// [PostgreSQL] unique index nulls handling option: `[ NULLS [ NOT ] DISTINCT ]`
 ///
-/// [Postgres]: https://www.postgresql.org/docs/17/sql-altertable.html
+/// [PostgreSQL]: https://www.postgresql.org/docs/17/sql-altertable.html
 #[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
@@ -2175,15 +2175,15 @@ pub struct CreateFunction {
     ///
     /// IMMUTABLE | STABLE | VOLATILE
     ///
-    /// [Postgres](https://www.postgresql.org/docs/current/sql-createfunction.html)
+    /// [PostgreSQL](https://www.postgresql.org/docs/current/sql-createfunction.html)
     pub behavior: Option<FunctionBehavior>,
     /// CALLED ON NULL INPUT | RETURNS NULL ON NULL INPUT | STRICT
     ///
-    /// [Postgres](https://www.postgresql.org/docs/current/sql-createfunction.html)
+    /// [PostgreSQL](https://www.postgresql.org/docs/current/sql-createfunction.html)
     pub called_on_null: Option<FunctionCalledOnNull>,
     /// PARALLEL { UNSAFE | RESTRICTED | SAFE }
     ///
-    /// [Postgres](https://www.postgresql.org/docs/current/sql-createfunction.html)
+    /// [PostgreSQL](https://www.postgresql.org/docs/current/sql-createfunction.html)
     pub parallel: Option<FunctionParallel>,
     /// USING ... (Hive only)
     pub using: Option<CreateFunctionUsing>,

src/ast/ddl.rs

@@ -410,7 +410,7 @@ impl fmt::Display for Interval {
 
 /// A field definition within a struct
 ///
-/// [bigquery]: https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types#struct_type
+/// [BigQuery]: https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types#struct_type
 #[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
@@ -431,7 +431,7 @@ impl fmt::Display for StructField {
 
 /// A field definition within a union
 ///
-/// [duckdb]: https://duckdb.org/docs/sql/data_types/union.html
+/// [DuckDB]: https://duckdb.org/docs/sql/data_types/union.html
 #[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
@@ -448,7 +448,7 @@ impl fmt::Display for UnionField {
 
 /// A dictionary field within a dictionary.
 ///
-/// [duckdb]: https://duckdb.org/docs/sql/data_types/struct#creating-structs
+/// [DuckDB]: https://duckdb.org/docs/sql/data_types/struct#creating-structs
 #[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
@@ -479,7 +479,7 @@ impl Display for Map {
 
 /// A map field within a map.
 ///
-/// [duckdb]: https://duckdb.org/docs/sql/data_types/map.html#creating-maps
+/// [DuckDB]: https://duckdb.org/docs/sql/data_types/map.html#creating-maps
 #[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
@@ -2385,10 +2385,10 @@ impl fmt::Display for DeclareAssignment {
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
 pub enum DeclareType {
-    /// Cursor variable type. e.g. [Snowflake] [Postgres]
+    /// Cursor variable type. e.g. [Snowflake] [PostgreSQL]
     ///
     /// [Snowflake]: https://docs.snowflake.com/en/developer-guide/snowflake-scripting/cursors#declaring-a-cursor
-    /// [Postgres]: https://www.postgresql.org/docs/current/plpgsql-cursors.html
+    /// [PostgreSQL]: https://www.postgresql.org/docs/current/plpgsql-cursors.html
     Cursor,
 
     /// Result set variable type. [Snowflake]
@@ -2427,15 +2427,15 @@ impl fmt::Display for DeclareType {
 }
 
 /// A `DECLARE` statement.
-/// [Postgres] [Snowflake] [BigQuery]
+/// [PostgreSQL] [Snowflake] [BigQuery]
 ///
 /// Examples:
 /// ```sql
 /// DECLARE variable_name := 42
 /// DECLARE liahona CURSOR FOR SELECT * FROM films;
 /// ```
 ///
-/// [Postgres]: https://www.postgresql.org/docs/current/sql-declare.html
+/// [PostgreSQL]: https://www.postgresql.org/docs/current/sql-declare.html
 /// [Snowflake]: https://docs.snowflake.com/en/sql-reference/snowflake-scripting/declare
 /// [BigQuery]: https://cloud.google.com/bigquery/docs/reference/standard-sql/procedural-language#declare
 #[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
@@ -3020,7 +3020,7 @@ pub enum Statement {
     /// ```sql
     /// CREATE ROLE
     /// ```
-    /// See [postgres](https://www.postgresql.org/docs/current/sql-createrole.html)
+    /// See [PostgreSQL](https://www.postgresql.org/docs/current/sql-createrole.html)
     CreateRole {
         names: Vec<ObjectName>,
         if_not_exists: bool,
@@ -3046,7 +3046,7 @@ pub enum Statement {
     /// ```sql
     /// CREATE SECRET
     /// ```
-    /// See [duckdb](https://duckdb.org/docs/sql/statements/create_secret.html)
+    /// See [DuckDB](https://duckdb.org/docs/sql/statements/create_secret.html)
     CreateSecret {
         or_replace: bool,
         temporary: Option<bool>,
@@ -3550,7 +3550,7 @@ pub enum Statement {
     ///
     /// Supported variants:
     /// 1. [Hive](https://cwiki.apache.org/confluence/display/hive/languagemanual+ddl#LanguageManualDDL-Create/Drop/ReloadFunction)
-    /// 2. [Postgres](https://www.postgresql.org/docs/15/sql-createfunction.html)
+    /// 2. [PostgreSQL](https://www.postgresql.org/docs/15/sql-createfunction.html)
     /// 3. [BigQuery](https://cloud.google.com/bigquery/docs/reference/standard-sql/data-definition-language#create_function_statement)
     CreateFunction(CreateFunction),
     /// CREATE TRIGGER
@@ -8281,7 +8281,7 @@ impl fmt::Display for FunctionDeterminismSpecifier {
 /// where within the statement, the body shows up.
 ///
 /// [BigQuery]: https://cloud.google.com/bigquery/docs/reference/standard-sql/data-definition-language#syntax_11
-/// [Postgres]: https://www.postgresql.org/docs/15/sql-createfunction.html
+/// [PostgreSQL]: https://www.postgresql.org/docs/15/sql-createfunction.html
 #[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
@@ -8319,7 +8319,7 @@ pub enum CreateFunctionBody {
     /// RETURN a + b;
     /// ```
     ///
-    /// [Postgres]: https://www.postgresql.org/docs/current/sql-createfunction.html
+    /// [PostgreSQL]: https://www.postgresql.org/docs/current/sql-createfunction.html
     Return(Expr),
 }
 
@@ -8625,9 +8625,9 @@ impl Display for CreateViewParams {
     }
 }
 
-/// Engine of DB. Some warehouse has parameters of engine, e.g. [clickhouse]
+/// Engine of DB. Some warehouse has parameters of engine, e.g. [ClickHouse]
 ///
-/// [clickhouse]: https://clickhouse.com/docs/en/engines/table-engines
+/// [ClickHouse]: https://clickhouse.com/docs/en/engines/table-engines
 #[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]

src/ast/mod.rs

@@ -995,7 +995,7 @@ pub trait Dialect: Debug + Any {
     /// Returns true if the dialect supports `SET NAMES <charset_name> [COLLATE <collation_name>]`.
     ///
     /// - [MySQL](https://dev.mysql.com/doc/refman/8.4/en/set-names.html)
-    /// - [Postgres](https://www.postgresql.org/docs/17/sql-set.html)
+    /// - [PostgreSQL](https://www.postgresql.org/docs/17/sql-set.html)
     ///
     /// Note: Postgres doesn't support the `COLLATE` clause, but we permissively parse it anyway.
     fn supports_set_names(&self) -> bool {

src/dialect/mod.rs

@@ -411,6 +411,7 @@ define_keywords!(
     HOSTS,
     HOUR,
     HOURS,
+    HUGEINT,
     ICEBERG,
     ID,
     IDENTITY,
@@ -908,7 +909,9 @@ define_keywords!(
     TRY_CONVERT,
     TUPLE,
     TYPE,
+    UBIGINT,
     UESCAPE,
+    UHUGEINT,
     UINT128,
     UINT16,
     UINT256,
@@ -942,6 +945,8 @@ define_keywords!(
     USER,
     USER_RESOURCES,
     USING,
+    USMALLINT,
+    UTINYINT,
     UUID,
     VACUUM,
     VALID,

src/keywords.rs

@@ -3779,7 +3779,7 @@ impl<'a> Parser<'a> {
         })
     }
 
-    /// Parse a postgresql casting style which is in the form of `expr::datatype`.
+    /// Parse a PostgreSQL casting style which is in the form of `expr::datatype`.
     pub fn parse_pg_cast(&mut self, expr: Expr) -> Result<Expr, ParserError> {
         Ok(Expr::Cast {
             kind: CastKind::DoubleColon,
@@ -4873,9 +4873,9 @@ impl<'a> Parser<'a> {
         }
     }
 
-    /// Parse `CREATE FUNCTION` for [Postgres]
+    /// Parse `CREATE FUNCTION` for [PostgreSQL]
     ///
-    /// [Postgres]: https://www.postgresql.org/docs/15/sql-createfunction.html
+    /// [PostgreSQL]: https://www.postgresql.org/docs/15/sql-createfunction.html
     fn parse_postgres_create_function(
         &mut self,
         or_replace: bool,
@@ -9171,6 +9171,11 @@ impl<'a> Parser<'a> {
                         Ok(DataType::BigInt(optional_precision?))
                     }
                 }
+                Keyword::HUGEINT => Ok(DataType::HugeInt),
+                Keyword::UBIGINT => Ok(DataType::UBigInt),
+                Keyword::UHUGEINT => Ok(DataType::UHugeInt),
+                Keyword::USMALLINT => Ok(DataType::USmallInt),
+                Keyword::UTINYINT => Ok(DataType::UTinyInt),
                 Keyword::UINT8 => Ok(DataType::UInt8),
                 Keyword::UINT16 => Ok(DataType::UInt16),
                 Keyword::UINT32 => Ok(DataType::UInt32),

src/parser/mod.rs

@@ -352,6 +352,32 @@ fn test_duckdb_load_extension() {
     );
 }
 
+#[test]
+fn test_duckdb_specific_int_types() {
+    let duckdb_dtypes = vec![
+        ("UTINYINT", DataType::UTinyInt),
+        ("USMALLINT", DataType::USmallInt),
+        ("UBIGINT", DataType::UBigInt),
+        ("UHUGEINT", DataType::UHugeInt),
+        ("HUGEINT", DataType::HugeInt),
+    ];
+    for (dtype_string, data_type) in duckdb_dtypes {
+        let sql = format!("SELECT 123::{}", dtype_string);
+        let select = duckdb().verified_only_select(&sql);
+        assert_eq!(
+            &Expr::Cast {
+                kind: CastKind::DoubleColon,
+                expr: Box::new(Expr::Value(
+                    Value::Number("123".parse().unwrap(), false).with_empty_span()
+                )),
+                data_type: data_type.clone(),
+                format: None,
+            },
+            expr_from_projection(&select.projection[0])
+        );
+    }
+}
+
 #[test]
 fn test_duckdb_struct_literal() {
     //struct literal syntax https://duckdb.org/docs/sql/data_types/struct#creating-structs