fix(sqlparser): typed string must be single quoted (#17482) (#17495)

Signed-off-by: Runji Wang <[email protected]>
Co-authored-by: xiangjinwu <[email protected]>
Co-authored-by: Runji Wang <[email protected]>

e2e_test/ddl/show.slt

@@ -102,12 +102,12 @@ show tables from public;
 t3
 
 query T
-show tables from public like "t_";
+show tables from public like 't_';
 ----
 t3
 
 query T
-show tables from public like "_t";
+show tables from public like '_t';
 ----
 
 query T

src/frontend/src/handler/create_sql_function.rs

@@ -150,6 +150,9 @@ pub async fn handle_create_sql_function(
     let body = match &params.as_ {
         Some(FunctionDefinition::SingleQuotedDef(s)) => s.clone(),
         Some(FunctionDefinition::DoubleDollarDef(s)) => s.clone(),
+        Some(FunctionDefinition::Identifier(_)) => {
+            return Err(ErrorCode::InvalidParameterValue("expect quoted string".to_string()).into())
+        }
         None => {
             if params.return_.is_none() {
                 return Err(ErrorCode::InvalidParameterValue(

src/sqlparser/src/ast/mod.rs

@@ -2885,13 +2885,15 @@ impl fmt::Display for FunctionBehavior {
 #[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 pub enum FunctionDefinition {
+    Identifier(String),
     SingleQuotedDef(String),
     DoubleDollarDef(String),
 }
 
 impl fmt::Display for FunctionDefinition {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         match self {
+            FunctionDefinition::Identifier(s) => write!(f, "{s}")?,
             FunctionDefinition::SingleQuotedDef(s) => write!(f, "'{s}'")?,
             FunctionDefinition::DoubleDollarDef(s) => write!(f, "$${s}$$")?,
         }
@@ -2903,6 +2905,7 @@ impl FunctionDefinition {
     /// Returns the function definition as a string slice.
     pub fn as_str(&self) -> &str {
         match self {
+            FunctionDefinition::Identifier(s) => s,
             FunctionDefinition::SingleQuotedDef(s) => s,
             FunctionDefinition::DoubleDollarDef(s) => s,
         }
@@ -2911,6 +2914,7 @@ impl FunctionDefinition {
     /// Returns the function definition as a string.
     pub fn into_string(self) -> String {
         match self {
+            FunctionDefinition::Identifier(s) => s,
             FunctionDefinition::SingleQuotedDef(s) => s,
             FunctionDefinition::DoubleDollarDef(s) => s,
         }

src/sqlparser/src/parser.rs

@@ -30,7 +30,9 @@ use winnow::{PResult, Parser as _};
 use crate::ast::*;
 use crate::keywords::{self, Keyword};
 use crate::parser_v2;
-use crate::parser_v2::{keyword, literal_i64, literal_uint, single_quoted_string, ParserExt as _};
+use crate::parser_v2::{
+    dollar_quoted_string, keyword, literal_i64, literal_uint, single_quoted_string, ParserExt as _,
+};
 use crate::tokenizer::*;
 
 pub(crate) const UPSTREAM_SOURCE_KEY: &str = "connector";
@@ -3550,28 +3552,20 @@ impl Parser<'_> {
     }
 
     pub fn parse_function_definition(&mut self) -> PResult<FunctionDefinition> {
-        let peek_token = self.peek_token();
-        match peek_token.token {
-            Token::DollarQuotedString(value) => {
-                self.next_token();
-                Ok(FunctionDefinition::DoubleDollarDef(value.value))
-            }
-            _ => Ok(FunctionDefinition::SingleQuotedDef(
-                self.parse_literal_string()?,
-            )),
-        }
+        alt((
+            single_quoted_string.map(FunctionDefinition::SingleQuotedDef),
+            dollar_quoted_string.map(FunctionDefinition::DoubleDollarDef),
+            Self::parse_identifier.map(|i| FunctionDefinition::Identifier(i.value)),
+            fail.expect("function definition"),
+        ))
+        .parse_next(self)
     }
 
     /// Parse a literal string
     pub fn parse_literal_string(&mut self) -> PResult<String> {
         let checkpoint = *self;
         let token = self.next_token();
         match token.token {
-            Token::Word(Word {
-                value,
-                keyword: Keyword::NoKeyword,
-                ..
-            }) => Ok(value),
             Token::SingleQuotedString(s) => Ok(s),
             _ => self.expected_at(checkpoint, "literal string"),
         }

src/sqlparser/src/parser_v2/mod.rs

@@ -129,6 +129,19 @@ where
         .parse_next(input)
 }
 
+/// Consume an $$ dollar-quoted string $$.
+pub fn dollar_quoted_string<S>(input: &mut S) -> PResult<String>
+where
+    S: TokenStream,
+{
+    token
+        .verify_map(|t| match &t.token {
+            Token::DollarQuotedString(s) => Some(s.value.clone()),
+            _ => None,
+        })
+        .parse_next(input)
+}
+
 /// Consume an object name.
 ///
 /// FIXME: Object name is extremely complex, we only handle a subset here.

src/sqlparser/tests/testdata/select.yaml

@@ -97,12 +97,10 @@
 - input: SELECT timestamp with time zone '2022-10-01 12:00:00Z' AT TIME ZONE zone
   formatted_sql: SELECT TIMESTAMP WITH TIME ZONE '2022-10-01 12:00:00Z' AT TIME ZONE zone
   formatted_ast: 'Query(Query { with: None, body: Select(Select { distinct: All, projection: [UnnamedExpr(AtTimeZone { timestamp: TypedString { data_type: Timestamp(true), value: "2022-10-01 12:00:00Z" }, time_zone: Identifier(Ident { value: "zone", quote_style: None }) })], from: [], lateral_views: [], selection: None, group_by: [], having: None }), order_by: [], limit: None, offset: None, fetch: None })'
-# https://www.postgresql.org/message-id/CADT4RqBPdbsZW7HS1jJP319TMRHs1hzUiP=iRJYR6UqgHCrgNQ@mail.gmail.com
-- input: SELECT now() + INTERVAL '14 days' AT TIME ZONE 'UTC';
+- input: SELECT now() + INTERVAL '14 days' AT TIME ZONE 'UTC'; -- https://www.postgresql.org/message-id/CADT4RqBPdbsZW7HS1jJP319TMRHs1hzUiP=iRJYR6UqgHCrgNQ@mail.gmail.com
   formatted_sql: SELECT now() + INTERVAL '14 days' AT TIME ZONE 'UTC'
   formatted_ast: 'Query(Query { with: None, body: Select(Select { distinct: All, projection: [UnnamedExpr(BinaryOp { left: Function(Function { name: ObjectName([Ident { value: "now", quote_style: None }]), args: [], variadic: false, over: None, distinct: false, order_by: [], filter: None, within_group: None }), op: Plus, right: AtTimeZone { timestamp: Value(Interval { value: "14 days", leading_field: None, leading_precision: None, last_field: None, fractional_seconds_precision: None }), time_zone: Value(SingleQuotedString("UTC")) } })], from: [], lateral_views: [], selection: None, group_by: [], having: None }), order_by: [], limit: None, offset: None, fetch: None })'
-# https://github.com/sqlparser-rs/sqlparser-rs/issues/1266
-- input: SELECT c FROM t WHERE c >= '2019-03-27T22:00:00.000Z'::timestamp AT TIME ZONE 'Europe/Brussels';
+- input: SELECT c FROM t WHERE c >= '2019-03-27T22:00:00.000Z'::timestamp AT TIME ZONE 'Europe/Brussels'; -- https://github.com/sqlparser-rs/sqlparser-rs/issues/1266
   formatted_sql: SELECT c FROM t WHERE c >= CAST('2019-03-27T22:00:00.000Z' AS TIMESTAMP) AT TIME ZONE 'Europe/Brussels'
   formatted_ast: 'Query(Query { with: None, body: Select(Select { distinct: All, projection: [UnnamedExpr(Identifier(Ident { value: "c", quote_style: None }))], from: [TableWithJoins { relation: Table { name: ObjectName([Ident { value: "t", quote_style: None }]), alias: None, as_of: None }, joins: [] }], lateral_views: [], selection: Some(BinaryOp { left: Identifier(Ident { value: "c", quote_style: None }), op: GtEq, right: AtTimeZone { timestamp: Cast { expr: Value(SingleQuotedString("2019-03-27T22:00:00.000Z")), data_type: Timestamp(false) }, time_zone: Value(SingleQuotedString("Europe/Brussels")) } }), group_by: [], having: None }), order_by: [], limit: None, offset: None, fetch: None })'
 - input: SELECT 0c6
@@ -222,3 +220,14 @@
     sql parser error: expected statement, found: selet
     LINE 1: selet 1;
             ^
+- input: select date t::date; -- https://github.com/risingwavelabs/risingwave/issues/17461
+  error_msg: |-
+    sql parser error: expected end of statement, found: ::
+    LINE 1: select date t::date; -- https://github.com/risingwavelabs/risingwave/issues/17461
+                         ^
+- input: select date 't'::date; -- TypedString higher precedence than Cast
+  formatted_sql: SELECT CAST(DATE 't' AS DATE)
+  formatted_ast: 'Query(Query { with: None, body: Select(Select { distinct: All, projection: [UnnamedExpr(Cast { expr: TypedString { data_type: Date, value: "t" }, data_type: Date })], from: [], lateral_views: [], selection: None, group_by: [], having: None }), order_by: [], limit: None, offset: None, fetch: None })'
+- input: select date t; -- A column "date" aliased to "t"
+  formatted_sql: SELECT date AS t
+  formatted_ast: 'Query(Query { with: None, body: Select(Select { distinct: All, projection: [ExprWithAlias { expr: Identifier(Ident { value: "date", quote_style: None }), alias: Ident { value: "t", quote_style: None } }], from: [], lateral_views: [], selection: None, group_by: [], having: None }), order_by: [], limit: None, offset: None, fetch: None })'

src/sqlparser/tests/testdata/show.yaml

@@ -11,7 +11,7 @@
 - input: SHOW TABLES FROM t
   formatted_sql: SHOW TABLES FROM t
   formatted_ast: 'ShowObjects { object: Table { schema: Some(Ident { value: "t", quote_style: None }) }, filter: None }'
-- input: SHOW TABLES FROM t LIKE "t%"
+- input: SHOW TABLES FROM t LIKE 't%'
   formatted_sql: SHOW TABLES FROM t LIKE 't%'
   formatted_ast: 'ShowObjects { object: Table { schema: Some(Ident { value: "t", quote_style: None }) }, filter: Some(Like("t%")) }'
 - input: SHOW VIEWS
@@ -29,7 +29,7 @@
 - input: SHOW INTERNAL TABLES FROM t
   formatted_sql: SHOW INTERNAL TABLES FROM t
   formatted_ast: 'ShowObjects { object: InternalTable { schema: Some(Ident { value: "t", quote_style: None }) }, filter: None }'
-- input: SHOW INTERNAL TABLES LIKE "%mv1%"
+- input: SHOW INTERNAL TABLES LIKE '%mv1%'
   formatted_sql: SHOW INTERNAL TABLES LIKE '%mv1%'
   formatted_ast: 'ShowObjects { object: InternalTable { schema: None }, filter: Some(Like("%mv1%")) }'
 - input: SHOW MATERIALIZED VIEWS FROM t