Add XMLPARSE expression

Author: LucaCappelletti94

src/ast/mod.rs

@@ -844,6 +844,43 @@ impl fmt::Display for CaseWhen {
     }
 }
 
+/// Parsing mode for `XMLPARSE`.
+#[derive(Debug, Clone, Copy, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub enum XmlParseMode {
+    /// `CONTENT`
+    Content,
+    /// `DOCUMENT`
+    Document,
+}
+
+impl fmt::Display for XmlParseMode {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match self {
+            XmlParseMode::Content => write!(f, "CONTENT"),
+            XmlParseMode::Document => write!(f, "DOCUMENT"),
+        }
+    }
+}
+
+/// `XMLPARSE(CONTENT|DOCUMENT expr)`.
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub struct XmlParseExpr {
+    /// Parsing mode.
+    pub mode: XmlParseMode,
+    /// Expression to parse as XML.
+    pub expr: Box<Expr>,
+}
+
+impl fmt::Display for XmlParseExpr {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "XMLPARSE({} {})", self.mode, self.expr)
+    }
+}
+
 /// An SQL expression of any type.
 ///
 /// # Semantics / Type Checking
@@ -1233,6 +1270,8 @@ pub enum Expr {
     /// This can represent ANSI SQL `DATE`, `TIME`, and `TIMESTAMP` literals (such as `DATE '2020-01-01'`),
     /// as well as constants of other types (a non-standard PostgreSQL extension).
     TypedString(TypedString),
+    /// XML parse expression: `XMLPARSE(CONTENT|DOCUMENT expr)`.
+    XmlParse(XmlParseExpr),
     /// Scalar function call e.g. `LEFT(foo, 5)`
     Function(Function),
     /// `CASE [<operand>] WHEN <condition> THEN <result> ... [ELSE <result>] END`
@@ -2015,6 +2054,7 @@ impl fmt::Display for Expr {
             Expr::Value(v) => write!(f, "{v}"),
             Expr::Prefixed { prefix, value } => write!(f, "{prefix} {value}"),
             Expr::TypedString(ts) => ts.fmt(f),
+            Expr::XmlParse(xml_parse) => xml_parse.fmt(f),
             Expr::Function(fun) => fun.fmt(f),
             Expr::Case {
                 case_token: _,

src/ast/spans.rs

@@ -1555,6 +1555,7 @@ impl Spanned for Expr {
             Expr::Nested(expr) => expr.span(),
             Expr::Value(value) => value.span(),
             Expr::TypedString(TypedString { value, .. }) => value.span(),
+            Expr::XmlParse(xml_parse) => xml_parse.expr.span(),
             Expr::Function(function) => function.span(),
             Expr::GroupingSets(vec) => {
                 union_spans(vec.iter().flat_map(|i| i.iter().map(|k| k.span())))

src/parser/mod.rs

@@ -2527,8 +2527,43 @@ impl<'a> Parser<'a> {
         })
     }
 
+    /// Consume the next token if it is an unquoted word matching `expected`
+    /// (case-insensitive), returning whether it was consumed.
+    fn parse_unquoted_word_value(&mut self, expected: &str) -> bool {
+        if let Token::Word(word) = &self.peek_token_ref().token {
+            if word.quote_style.is_none() && word.value.eq_ignore_ascii_case(expected) {
+                self.next_token();
+                return true;
+            }
+        }
+        false
+    }
+
+    fn parse_xml_parse_mode(&mut self) -> Result<XmlParseMode, ParserError> {
+        if self.parse_unquoted_word_value("content") {
+            Ok(XmlParseMode::Content)
+        } else if self.parse_unquoted_word_value("document") {
+            Ok(XmlParseMode::Document)
+        } else {
+            self.expected_ref("CONTENT or DOCUMENT", self.peek_token_ref())
+        }
+    }
+
+    fn parse_xmlparse_expr(&mut self) -> Result<Expr, ParserError> {
+        self.expect_token(&Token::LParen)?;
+        let mode = self.parse_xml_parse_mode()?;
+        let expr = Box::new(self.parse_expr()?);
+        self.expect_token(&Token::RParen)?;
+        Ok(Expr::XmlParse(XmlParseExpr { mode, expr }))
+    }
+
     /// Parse a function call expression named by `name` and return it as an `Expr`.
     pub fn parse_function(&mut self, name: ObjectName) -> Result<Expr, ParserError> {
+        if self.dialect.supports_xml_expressions()
+            && Self::is_simple_unquoted_object_name(&name, "xmlparse")
+        {
+            return self.parse_xmlparse_expr();
+        }
         self.parse_function_call(name).map(Expr::Function)
     }
 

tests/sqlparser_common.rs

@@ -19014,6 +19014,46 @@ fn parse_aliased_function_args() {
         .is_err());
 }
 
+#[test]
+fn parse_xmlparse() {
+    let dialects = all_dialects_where(|d| d.supports_xml_expressions());
+
+    let select = dialects.verified_only_select_with_canonical(
+        "SELECT xmlparse(content '<a/>')",
+        "SELECT XMLPARSE(CONTENT '<a/>')",
+    );
+    match &select.projection[0] {
+        UnnamedExpr(Expr::XmlParse(XmlParseExpr { mode, .. })) => {
+            assert_eq!(*mode, XmlParseMode::Content);
+        }
+        item => panic!("expected XmlParse expression, got {item:?}"),
+    }
+
+    let select = dialects.verified_only_select_with_canonical(
+        "SELECT xmlparse(document '<a/>')",
+        "SELECT XMLPARSE(DOCUMENT '<a/>')",
+    );
+    match &select.projection[0] {
+        UnnamedExpr(Expr::XmlParse(XmlParseExpr { mode, .. })) => {
+            assert_eq!(*mode, XmlParseMode::Document);
+        }
+        item => panic!("expected XmlParse expression, got {item:?}"),
+    }
+
+    // XMLPARSE requires a CONTENT or DOCUMENT mode.
+    assert!(dialects
+        .parse_sql_statements("SELECT xmlparse('<a/>')")
+        .is_err());
+
+    // On dialects without XML support, `xmlparse` stays a regular function
+    // and the special `CONTENT <expr>` syntax is rejected.
+    let others = all_dialects_except(|d| d.supports_xml_expressions());
+    others.verified_only_select("SELECT xmlparse(1)");
+    assert!(others
+        .parse_sql_statements("SELECT xmlparse(content '<a/>')")
+        .is_err());
+}
+
 /// Regression test for the 2^N parse-time blowup in `parse_compound_expr` on
 /// inputs like `IF a0.a1...aN.#`. The parse is run on a worker thread and the
 /// main thread asserts that it reports back within a generous timeout. Post-fix

tests/sqlparser_postgres.rs

@@ -3951,6 +3951,47 @@ fn parse_xmlforest_aliased_arguments() {
     );
 }
 
+#[test]
+fn parse_xmlparse() {
+    // Regression statements covering CONTENT and DOCUMENT modes with valid,
+    // invalid, and edge-case XML strings (parsing only, no semantic checks).
+    let statements = [
+        "SELECT XMLPARSE(CONTENT '')",
+        "SELECT XMLPARSE(CONTENT ' ')",
+        "SELECT XMLPARSE(CONTENT 'abc')",
+        "SELECT XMLPARSE(CONTENT '<abc>x</abc>')",
+        "SELECT XMLPARSE(CONTENT '<invalidentity>&</invalidentity>')",
+        "SELECT XMLPARSE(CONTENT '<undefinedentity>&idontexist;</undefinedentity>')",
+        "SELECT XMLPARSE(CONTENT '<twoerrors>&idontexist;</unbalanced>')",
+        "SELECT XMLPARSE(CONTENT '<nosuchprefix:tag/>')",
+        "SELECT XMLPARSE(DOCUMENT ' ')",
+        "SELECT XMLPARSE(DOCUMENT 'abc')",
+        "SELECT XMLPARSE(DOCUMENT '<abc>x</abc>')",
+        "SELECT XMLPARSE(DOCUMENT '<invalidentity>&</abc>')",
+        "SELECT XMLPARSE(DOCUMENT '<undefinedentity>&idontexist;</abc>')",
+        "SELECT XMLPARSE(DOCUMENT '<twoerrors>&idontexist;</unbalanced>')",
+        "SELECT XMLPARSE(DOCUMENT '<nosuchprefix:tag/>')",
+    ];
+    for sql in statements {
+        pg().verified_stmt(sql);
+    }
+
+    // Lowercase keywords canonicalize to uppercase.
+    let select = pg().verified_only_select_with_canonical(
+        "SELECT xmlparse(content '<a/>')",
+        "SELECT XMLPARSE(CONTENT '<a/>')",
+    );
+    assert_eq!(
+        expr_from_projection(&select.projection[0]),
+        &Expr::XmlParse(XmlParseExpr {
+            mode: XmlParseMode::Content,
+            expr: Box::new(Expr::Value(
+                Value::SingleQuotedString("<a/>".to_string()).into()
+            )),
+        })
+    );
+}
+
 #[test]
 fn parse_xml_typed_string() {
     // xml '...' should parse as a TypedString on PostgreSQL and Generic