Support use of BY NAME quantifier across all set ops (#1309)

Co-authored-by: Alexander Beedie <alexander.beedie@adia.ae>
Co-authored-by: Joey Hain <joey@sigmacomputing.com>

README.md

@@ -114,13 +114,12 @@ $ cargo run --features json_example --example cli FILENAME.sql [--dialectname]
 
 ## Users
 
-This parser is currently being used by the [DataFusion] query engine,
+This parser is currently being used by the [DataFusion] query engine, [LocustDB],
-[LocustDB], [Ballista], [GlueSQL], [Opteryx], [PRQL], [Qrlew], [JumpWire], and [ParadeDB].
+[Ballista], [GlueSQL], [Opteryx], [Polars], [PRQL], [Qrlew], [JumpWire], and [ParadeDB].
 
 If your project is using sqlparser-rs feel free to make a PR to add it
 to this list.
 
-
 ## Design
 
 The core expression parser uses the [Pratt Parser] design, which is a top-down
@@ -210,6 +209,7 @@ licensed as above, without any additional terms or conditions.
 [Ballista]: https://github.com/apache/arrow-ballista
 [GlueSQL]: https://github.com/gluesql/gluesql
 [Opteryx]: https://github.com/mabel-dev/opteryx
+[Polars]: https://pola.rs/
 [PRQL]: https://github.com/PRQL/prql
 [Qrlew]: https://github.com/Qrlew/qrlew
 [JumpWire]: https://github.com/extragoodlabs/jumpwire

src/ast/data_type.rs

@@ -168,7 +168,7 @@ pub enum DataType {
     UnsignedInt(Option<u64>),
     /// Unsigned int4 with optional display width e.g. INT4 UNSIGNED or INT4(11) UNSIGNED
     UnsignedInt4(Option<u64>),
-    /// Unsigned integer with optional display width e.g. INTGER UNSIGNED or INTEGER(11) UNSIGNED
+    /// Unsigned integer with optional display width e.g. INTEGER UNSIGNED or INTEGER(11) UNSIGNED
     UnsignedInteger(Option<u64>),
     /// Unsigned integer type in [clickhouse]
     /// Note: UInt8 mean 8 bits in [clickhouse]
@@ -699,7 +699,7 @@ pub enum CharacterLength {
         /// Optional unit. If not informed, the ANSI handles it as CHARACTERS implicitly
         unit: Option<CharLengthUnits>,
     },
-    /// VARCHAR(MAX) or NVARCHAR(MAX), used in T-SQL (Miscrosoft SQL Server)
+    /// VARCHAR(MAX) or NVARCHAR(MAX), used in T-SQL (Microsoft SQL Server)
     Max,
 }
 

src/ast/mod.rs

@@ -2265,7 +2265,7 @@ pub enum Statement {
     /// SET [ SESSION | LOCAL ] ROLE role_name
     /// ```
     ///
-    /// Sets sesssion state. Examples: [ANSI][1], [Postgresql][2], [MySQL][3], and [Oracle][4]
+    /// Sets session state. Examples: [ANSI][1], [Postgresql][2], [MySQL][3], and [Oracle][4]
     ///
     /// [1]: https://jakewheat.github.io/sql-overview/sql-2016-foundation-grammar.html#set-role-statement
     /// [2]: https://www.postgresql.org/docs/14/sql-set-role.html
@@ -2283,7 +2283,7 @@ pub enum Statement {
     /// ```
     ///
     /// Note: this is not a standard SQL statement, but it is supported by at
-    /// least MySQL and PostgreSQL. Not all MySQL-specific syntatic forms are
+    /// least MySQL and PostgreSQL. Not all MySQL-specific syntactic forms are
     /// supported yet.
     SetVariable {
         local: bool,
@@ -4750,7 +4750,7 @@ impl fmt::Display for FunctionArguments {
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
 pub struct FunctionArgumentList {
-    /// `[ ALL | DISTINCT ]
+    /// `[ ALL | DISTINCT ]`
     pub duplicate_treatment: Option<DuplicateTreatment>,
     /// The function arguments.
     pub args: Vec<FunctionArg>,

src/parser/mod.rs

@@ -8138,7 +8138,7 @@ impl<'a> Parser<'a> {
 
     pub fn parse_set_quantifier(&mut self, op: &Option<SetOperator>) -> SetQuantifier {
         match op {
-            Some(SetOperator::Union) => {
+            Some(SetOperator::Except | SetOperator::Intersect | SetOperator::Union) => {
                 if self.parse_keywords(&[Keyword::DISTINCT, Keyword::BY, Keyword::NAME]) {
                     SetQuantifier::DistinctByName
                 } else if self.parse_keywords(&[Keyword::BY, Keyword::NAME]) {
@@ -8155,15 +8155,6 @@ impl<'a> Parser<'a> {
                     SetQuantifier::None
                 }
             }
-            Some(SetOperator::Except) | Some(SetOperator::Intersect) => {
-                if self.parse_keyword(Keyword::ALL) {
-                    SetQuantifier::All
-                } else if self.parse_keyword(Keyword::DISTINCT) {
-                    SetQuantifier::Distinct
-                } else {
-                    SetQuantifier::None
-                }
-            }
             _ => SetQuantifier::None,
         }
     }
@@ -8547,10 +8538,10 @@ impl<'a> Parser<'a> {
             })
         } else if variable.to_string() == "TRANSACTION" && modifier.is_none() {
             if self.parse_keyword(Keyword::SNAPSHOT) {
-                let snaphot_id = self.parse_value()?;
+                let snapshot_id = self.parse_value()?;
                 return Ok(Statement::SetTransaction {
                     modes: vec![],
-                    snapshot: Some(snaphot_id),
+                    snapshot: Some(snapshot_id),
                     session: false,
                 });
             }

src/tokenizer.rs

@@ -654,7 +654,7 @@ impl<'a> Tokenizer<'a> {
         Ok(())
     }
 
-    // Tokenize the identifer or keywords in `ch`
+    // Tokenize the identifier or keywords in `ch`
     fn tokenize_identifier_or_keyword(
         &self,
         ch: impl IntoIterator<Item = char>,

tests/sqlparser_common.rs

@@ -6010,6 +6010,12 @@ fn parse_union_except_intersect() {
     verified_stmt("SELECT foo FROM tab UNION SELECT bar FROM TAB");
     verified_stmt("(SELECT * FROM new EXCEPT SELECT * FROM old) UNION ALL (SELECT * FROM old EXCEPT SELECT * FROM new) ORDER BY 1");
     verified_stmt("(SELECT * FROM new EXCEPT DISTINCT SELECT * FROM old) UNION DISTINCT (SELECT * FROM old EXCEPT DISTINCT SELECT * FROM new) ORDER BY 1");
+    verified_stmt("SELECT 1 AS x, 2 AS y EXCEPT BY NAME SELECT 9 AS y, 8 AS x");
+    verified_stmt("SELECT 1 AS x, 2 AS y EXCEPT ALL BY NAME SELECT 9 AS y, 8 AS x");
+    verified_stmt("SELECT 1 AS x, 2 AS y EXCEPT DISTINCT BY NAME SELECT 9 AS y, 8 AS x");
+    verified_stmt("SELECT 1 AS x, 2 AS y INTERSECT BY NAME SELECT 9 AS y, 8 AS x");
+    verified_stmt("SELECT 1 AS x, 2 AS y INTERSECT ALL BY NAME SELECT 9 AS y, 8 AS x");
+    verified_stmt("SELECT 1 AS x, 2 AS y INTERSECT DISTINCT BY NAME SELECT 9 AS y, 8 AS x");
 }
 
 #[test]