fix: share MySQL character set parsing across column types and casts (#2662)

Signed-off-by: minleejae <mmj9808@gmail.com>

Author: minleejae

src/main/java/net/sf/jsqlparser/statement/create/table/ColDataType.java

@@ -40,6 +40,7 @@ public enum NationalCharacterType {
     private String dataType;
     private List<String> argumentsStringList;
     private String characterSet;
+    private boolean useCharsetKeyword;
     private IntervalQualifier intervalQualifier;
     private List<Integer> arrayData = new ArrayList<Integer>();
     private Signedness signedness;
@@ -173,6 +174,15 @@ public void setCharacterSet(String characterSet) {
         this.characterSet = characterSet;
     }
 
+    /** Whether the character set clause uses MySQL's CHARSET abbreviation. */
+    public boolean isUseCharsetKeyword() {
+        return useCharsetKeyword;
+    }
+
+    public void setUseCharsetKeyword(boolean useCharsetKeyword) {
+        this.useCharsetKeyword = useCharsetKeyword;
+    }
+
     public IntervalQualifier getIntervalQualifier() {
         return intervalQualifier;
     }
@@ -338,7 +348,9 @@ public String toString() {
                         : (signedness != null ? " " + signedness : "")
                                 + (zerofill ? " ZEROFILL" : ""))
                 + arraySpec.toString()
-                + (characterSet != null ? " CHARACTER SET " + characterSet : "");
+                + (characterSet != null
+                        ? (useCharsetKeyword ? " CHARSET " : " CHARACTER SET ") + characterSet
+                        : "");
     }
 
     public ColDataType withDataType(String dataType) {
@@ -435,6 +447,7 @@ public final boolean equals(Object o) {
         return dataType.equalsIgnoreCase(that.dataType)
                 && Objects.equals(argumentsStringList, that.argumentsStringList)
                 && Objects.equals(characterSet, that.characterSet)
+                && useCharsetKeyword == that.useCharsetKeyword
                 && Objects.equals(intervalQualifier, that.intervalQualifier)
                 && Objects.equals(arrayData, that.arrayData)
                 && signedness == that.signedness
@@ -452,6 +465,7 @@ public int hashCode() {
                 .reduce(0, (hash, c) -> 31 * hash + c);
         result = 31 * result + Objects.hashCode(argumentsStringList);
         result = 31 * result + Objects.hashCode(characterSet);
+        result = 31 * result + Boolean.hashCode(useCharsetKeyword);
         result = 31 * result + Objects.hashCode(intervalQualifier);
         result = 31 * result + Objects.hashCode(arrayData);
         result = 31 * result + Objects.hashCode(signedness);

src/main/jjtree/net/sf/jsqlparser/parser/JSqlParserCC.jjt

@@ -15358,7 +15358,8 @@ ColDataType ColDataType():
     ( LOOKAHEAD(1) typeModifier=MySqlTypeModifier()
         { colDataType.addTypeModifier(typeModifier); } )*
     [ LOOKAHEAD(2) ( LOOKAHEAD(2) "[" {tk=null;} [ tk=<S_LONG> ] { array.add(tk!=null?Integer.valueOf(tk.image):null); } "]" )+ { colDataType.setArrayData(array); } ]
-    [ LOOKAHEAD(2) <K_CHARACTER> <K_SET> (tk=<S_IDENTIFIER> | tk=<K_BINARY>) { colDataType.setCharacterSet(tk.image); } ]
+    [ LOOKAHEAD({ (getToken(1).kind == K_CHARACTER && getToken(2).kind == K_SET)
+            || isKeywordAhead("CHARSET") }) TypeCharacterSet(colDataType) ]
 
     {
         requireDdlSyntax(colDataType.getXmlTypeModifier() == null || argumentsStringList.isEmpty(),
@@ -15381,6 +15382,19 @@ ColDataType ColDataType():
     }
 }
 
+/** Shared by column definitions, casts and other uses of ColDataType. */
+void TypeCharacterSet(ColDataType type):
+{
+    Token name;
+    boolean abbreviated = false;
+}
+{
+    ( <K_CHARACTER> <K_SET>
+      | LOOKAHEAD({ isKeywordAhead("CHARSET") }) <S_IDENTIFIER> { abbreviated = true; } )
+    ( name=<S_IDENTIFIER> | name=<K_BINARY> | name=<S_CHAR_LITERAL> | name=<S_QUOTED_IDENTIFIER> )
+    { type.setCharacterSet(name.image); type.setUseCharsetKeyword(abbreviated); }
+}
+
 XmlTypeModifier XmlTypeModifier():
 {
     XmlTypeModifier.Kind kind = null;

src/test/java/net/sf/jsqlparser/statement/create/table/MySqlTypeCharacterSetTest.java

@@ -0,0 +1,83 @@
+/*-
+ * #%L
+ * JSQLParser library
+ * %%
+ * Copyright (C) 2004 - 2026 JSQLParser
+ * %%
+ * Dual licensed under GNU LGPL 2.1 or Apache License 2.0
+ * #L%
+ */
+package net.sf.jsqlparser.statement.create.table;
+
+import net.sf.jsqlparser.expression.CastExpression;
+import net.sf.jsqlparser.parser.AbstractJSqlParser.Dialect;
+import net.sf.jsqlparser.parser.CCJSqlParserUtil;
+import net.sf.jsqlparser.statement.Statement;
+import net.sf.jsqlparser.statement.select.PlainSelect;
+import net.sf.jsqlparser.util.deparser.StatementDeParser;
+import org.junit.jupiter.api.Test;
+import org.junit.jupiter.params.ParameterizedTest;
+import org.junit.jupiter.params.provider.ValueSource;
+
+import static org.junit.jupiter.api.Assertions.*;
+
+class MySqlTypeCharacterSetTest {
+    private Statement parse(String sql) throws Exception {
+        return CCJSqlParserUtil.parse(sql, p -> p.withDialect(Dialect.MYSQL));
+    }
+
+    @ParameterizedTest
+    @ValueSource(strings = {
+            "SELECT CAST('xx' AS CHAR(16) CHARSET BINARY)",
+            "SELECT CAST('xx' AS CHAR(16) CHARACTER SET 'utf8mb4')",
+            "SELECT CONVERT('xx' USING utf8mb4)",
+            "CREATE TABLE t (c VARCHAR(20) CHARACTER SET 'utf8mb4' COLLATE 'utf8mb4_bin')",
+            "CREATE TABLE t (c TEXT CHARSET utf8mb4)",
+            "CREATE TABLE t (c SET('a','b') CHARSET 'utf8mb4')",
+            "CREATE TABLE t (c BINARY(16) AS (CAST('xx' AS CHAR(16) CHARSET BINARY)))",
+            "ALTER TABLE t MODIFY c ENUM('a','b') CHARACTER SET 'utf8' COLLATE 'utf8_unicode_ci'",
+            "ALTER TABLE t ADD c CHAR(4) CHARSET utf8mb4, ADD d VARCHAR(10) CHARACTER SET utf8mb4"
+    })
+    void sharesTypeSyntaxAcrossCastsAndColumnDefinitions(String sql) throws Exception {
+        Statement statement = parse(sql);
+        StringBuilder output = new StringBuilder();
+        statement.accept(new StatementDeParser(output), null);
+        assertEquals(statement.toString(), output.toString());
+        assertEquals(statement.toString(), parse(output.toString()).toString());
+    }
+
+    @Test
+    void exposesCharsetAndRetainsAbbreviationWhenMutated() throws Exception {
+        CreateTable table = (CreateTable) parse("CREATE TABLE t (c VARCHAR(20) CHARSET 'utf8mb4')");
+        ColDataType type = table.getColumnDefinitions().get(0).getColDataType();
+        assertEquals("'utf8mb4'", type.getCharacterSet());
+        assertTrue(type.isUseCharsetKeyword());
+        type.setCharacterSet("latin1");
+        assertEquals("CREATE TABLE t (c VARCHAR (20) CHARSET latin1)", table.toString());
+        type.setUseCharsetKeyword(false);
+        assertEquals("CREATE TABLE t (c VARCHAR (20) CHARACTER SET latin1)", table.toString());
+        type.setCharacterSet(null);
+        assertEquals("CREATE TABLE t (c VARCHAR (20))", table.toString());
+    }
+
+    @Test
+    void preservesDefaultApiSpellingAndTypeEquality() throws Exception {
+        ColDataType legacy = new ColDataType("CHAR").withCharacterSet("binary");
+        assertEquals("CHAR CHARACTER SET binary", legacy.toString());
+        PlainSelect select = (PlainSelect) parse("SELECT CAST('a' AS CHAR CHARACTER SET binary)");
+        ColDataType parsed =
+                ((CastExpression) select.getSelectItem(0).getExpression()).getColDataType();
+        assertEquals(legacy, parsed);
+        assertEquals(legacy.hashCode(), parsed.hashCode());
+        parsed.setUseCharsetKeyword(true);
+        assertNotEquals(legacy, parsed);
+    }
+
+    @ParameterizedTest
+    @ValueSource(strings = {"SELECT CAST('x' AS CHAR CHARSET)",
+            "SELECT CAST('x' AS CHAR CHARACTER SET)",
+            "SELECT CAST('x' AS CHAR CHARSET = utf8mb4)"})
+    void requiresCharsetName(String sql) {
+        assertThrows(Exception.class, () -> parse(sql));
+    }
+}