Support MATCH AGAINST (apache#708)

Support added for both MySQL and Generic dialects.

Author: AugustoFKL

src/ast/mod.rs

@@ -449,6 +449,26 @@ pub enum Expr {
         /// or as `__ TO SECOND(x)`.
         fractional_seconds_precision: Option<u64>,
     },
+    /// `MySQL` specific text search function [(1)].
+    ///
+    /// Syntax:
+    /// ```text
+    /// MARCH (<col>, <col>, ...) AGAINST (<expr> [<search modifier>])
+    ///
+    /// <col> = CompoundIdentifier
+    /// <expr> = String literal
+    /// ```
+    ///
+    ///
+    /// [(1)]: https://dev.mysql.com/doc/refman/8.0/en/fulltext-search.html#function_match
+    MatchAgainst {
+        /// `(<col>, <col>, ...)`.
+        columns: Vec<Ident>,
+        /// `<expr>`.
+        match_value: Value,
+        /// `<search modifier>`
+        opt_search_modifier: Option<SearchModifier>,
+    },
 }
 
 impl fmt::Display for Expr {
@@ -818,6 +838,21 @@ impl fmt::Display for Expr {
                 }
                 Ok(())
             }
+            Expr::MatchAgainst {
+                columns,
+                match_value: match_expr,
+                opt_search_modifier,
+            } => {
+                write!(f, "MATCH ({}) AGAINST ", display_comma_separated(columns),)?;
+
+                if let Some(search_modifier) = opt_search_modifier {
+                    write!(f, "({match_expr} {search_modifier})")?;
+                } else {
+                    write!(f, "({match_expr})")?;
+                }
+
+                Ok(())
+            }
         }
     }
 }
@@ -3659,6 +3694,43 @@ impl fmt::Display for SchemaName {
     }
 }
 
+/// Fulltext search modifiers ([1]).
+///
+/// [1]: https://dev.mysql.com/doc/refman/8.0/en/fulltext-search.html#function_match
+#[derive(Debug, Clone, PartialEq, Eq, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+pub enum SearchModifier {
+    /// `IN NATURAL LANGUAGE MODE`.
+    InNaturalLanguageMode,
+    /// `IN NATURAL LANGUAGE MODE WITH QUERY EXPANSION`.
+    InNaturalLanguageModeWithQueryExpansion,
+    ///`IN BOOLEAN MODE`.
+    InBooleanMode,
+    ///`WITH QUERY EXPANSION`.
+    WithQueryExpansion,
+}
+
+impl fmt::Display for SearchModifier {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match self {
+            Self::InNaturalLanguageMode => {
+                write!(f, "IN NATURAL LANGUAGE MODE")?;
+            }
+            Self::InNaturalLanguageModeWithQueryExpansion => {
+                write!(f, "IN NATURAL LANGUAGE MODE WITH QUERY EXPANSION")?;
+            }
+            Self::InBooleanMode => {
+                write!(f, "IN BOOLEAN MODE")?;
+            }
+            Self::WithQueryExpansion => {
+                write!(f, "WITH QUERY EXPANSION")?;
+            }
+        }
+
+        Ok(())
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;

src/keywords.rs

@@ -71,6 +71,7 @@ define_keywords!(
     ACTION,
     ADD,
     ADMIN,
+    AGAINST,
     ALL,
     ALLOCATE,
     ALTER,
@@ -229,6 +230,7 @@ define_keywords!(
     EXECUTE,
     EXISTS,
     EXP,
+    EXPANSION,
     EXPLAIN,
     EXTENDED,
     EXTERNAL,
@@ -348,6 +350,7 @@ define_keywords!(
     MINUTE,
     MINVALUE,
     MOD,
+    MODE,
     MODIFIES,
     MODULE,
     MONTH,

src/parser.rs

@@ -505,6 +505,9 @@ impl<'a> Parser<'a> {
                 }
                 Keyword::ARRAY_AGG => self.parse_array_agg_expr(),
                 Keyword::NOT => self.parse_not(),
+                Keyword::MATCH if dialect_of!(self is MySqlDialect | GenericDialect) => {
+                    self.parse_match_against()
+                }
                 // Here `w` is a word, check if it's a part of a multi-part
                 // identifier, a function call, or a simple identifier:
                 _ => match self.peek_token() {
@@ -1209,6 +1212,57 @@ impl<'a> Parser<'a> {
         }
     }
 
+    /// Parses fulltext expressions [(1)]
+    ///
+    /// # Errors
+    /// This method will raise an error if the column list is empty or with invalid identifiers,
+    /// the match expression is not a literal string, or if the search modifier is not valid.
+    ///
+    /// [(1)]: Expr::MatchAgainst
+    pub fn parse_match_against(&mut self) -> Result<Expr, ParserError> {
+        let columns = self.parse_parenthesized_column_list(Mandatory)?;
+
+        self.expect_keyword(Keyword::AGAINST)?;
+
+        self.expect_token(&Token::LParen)?;
+
+        // MySQL is too permissive about the value, IMO we can't validate it perfectly on syntax level.
+        let match_value = self.parse_value()?;
+
+        let in_natural_language_mode_keywords = &[
+            Keyword::IN,
+            Keyword::NATURAL,
+            Keyword::LANGUAGE,
+            Keyword::MODE,
+        ];
+
+        let with_query_expansion_keywords = &[Keyword::WITH, Keyword::QUERY, Keyword::EXPANSION];
+
+        let in_boolean_mode_keywords = &[Keyword::IN, Keyword::BOOLEAN, Keyword::MODE];
+
+        let opt_search_modifier = if self.parse_keywords(in_natural_language_mode_keywords) {
+            if self.parse_keywords(with_query_expansion_keywords) {
+                Some(SearchModifier::InNaturalLanguageModeWithQueryExpansion)
+            } else {
+                Some(SearchModifier::InNaturalLanguageMode)
+            }
+        } else if self.parse_keywords(in_boolean_mode_keywords) {
+            Some(SearchModifier::InBooleanMode)
+        } else if self.parse_keywords(with_query_expansion_keywords) {
+            Some(SearchModifier::WithQueryExpansion)
+        } else {
+            None
+        };
+
+        self.expect_token(&Token::RParen)?;
+
+        Ok(Expr::MatchAgainst {
+            columns,
+            match_value,
+            opt_search_modifier,
+        })
+    }
+
     /// Parse an INTERVAL expression.
     ///
     /// Some syntactically valid intervals:

tests/sqlparser_mysql.rs

@@ -1166,6 +1166,32 @@ fn parse_create_table_with_spatial_definition() {
         .verified_stmt("CREATE TABLE tb (c1 INT, c2 INT, SPATIAL KEY potato (c1, c2))");
 }
 
+#[test]
+fn parse_fulltext_expression() {
+    mysql_and_generic().verified_stmt("SELECT * FROM tb WHERE MATCH (c1) AGAINST ('string')");
+
+    mysql_and_generic().verified_stmt(
+        "SELECT * FROM tb WHERE MATCH (c1) AGAINST ('string' IN NATURAL LANGUAGE MODE)",
+    );
+
+    mysql_and_generic().verified_stmt("SELECT * FROM tb WHERE MATCH (c1) AGAINST ('string' IN NATURAL LANGUAGE MODE WITH QUERY EXPANSION)");
+
+    mysql_and_generic()
+        .verified_stmt("SELECT * FROM tb WHERE MATCH (c1) AGAINST ('string' IN BOOLEAN MODE)");
+
+    mysql_and_generic()
+        .verified_stmt("SELECT * FROM tb WHERE MATCH (c1) AGAINST ('string' WITH QUERY EXPANSION)");
+
+    mysql_and_generic()
+        .verified_stmt("SELECT * FROM tb WHERE MATCH (c1, c2, c3) AGAINST ('string')");
+
+    mysql_and_generic().verified_stmt("SELECT * FROM tb WHERE MATCH (c1) AGAINST (123)");
+
+    mysql_and_generic().verified_stmt("SELECT * FROM tb WHERE MATCH (c1) AGAINST (NULL)");
+
+    mysql_and_generic().verified_stmt("SELECT COUNT(IF(MATCH (title, body) AGAINST ('database' IN NATURAL LANGUAGE MODE), 1, NULL)) AS count FROM articles");
+}
+
 #[test]
 #[should_panic = "Expected FULLTEXT or SPATIAL option without constraint name, found: cons"]
 fn parse_create_table_with_fulltext_definition_should_not_accept_constraint_name() {