java - 使用ANTLR4解析mysql简单例子

Question

我在这里使用 mysql 语法:https://github.com/antlr/grammars-v4/tree/master/mysql并使用 Maven 生成了 java 文件。现在，我试图解析一个查询，但不知道如何解析。

我基本上想“获取”查询的所有不同组成部分，例如所选的列表列、条件、子查询、表名等。但我不知道如何进行。到目前为止，我已经编写了以下代码。有人可以用一个简单的例子提出建议，以便我可以理解用法并处理更复杂的任务吗？这是我的代码:

public static void main( String[] args )
{
    String sql="select cust_name from database..table where cust_name like 'Kash%'";
    ANTLRInputStream input = new ANTLRInputStream(sql);
    MySqlLexer mySqlLexer = new MySqlLexer(input);
    CommonTokenStream tokens = new CommonTokenStream(mySqlLexer);
    MySqlParser mySqlParser = new MySqlParser(tokens);

    ParseTree tree = mySqlParser.dmlStatement();
    ParseTreeWalker walker = new ParseTreeWalker();
    MySqlParserBaseListener listener=new MySqlParserBaseListener();
    ParseTreeWalker.DEFAULT.walk(listener, tree);


    System.out.println(?);
}  

使用上面的代码，我得到以下输出:

line 1:11 no viable alternative at input '_'
(dmlStatement _ . . _ 'Kash%')

感谢帮助:)

Answer 1

I basically want to 'get' all the different components of a query, like the list columns selected, where conditions, sub queries, table names, etc.

您的 tree 变量保存所有数据:ParseTree tree = mySqlParser.dmlStatement();

line 1:11 no viable alternative at input '_'

如果您查看词法分析器规则:

SELECT:                              'SELECT';

ID:                                  ID_LITERAL;

fragment ID_LITERAL:                 [A-Z_$0-9]*?[A-Z_$]+?[A-Z_$0-9]*;

看来关键字和标识符不能包含小写字母。

如果你这样运行它:

String sql = "SELECT CUST_NAME FROM CUSTOMERS WHERE CUST_NAME LIKE 'Kash%'";

MySqlLexer lexer = new MySqlLexer(CharStreams.fromString(sql));
MySqlParser parser = new MySqlParser(new CommonTokenStream(lexer));
ParseTree root = parser.dmlStatement();

System.out.println(root.toStringTree(parser));

您将看到以下输出(为了便于阅读而缩进):

(dmlStatement 
  (selectStatement 
    (querySpecification SELECT 
      (selectElements 
        (selectElement 
          (fullColumnName 
            (uid 
              (simpleId CUST_NAME))))) 
      (fromClause FROM 
        (tableSources 
          (tableSource 
            (tableSourceItem 
              (tableName 
                (fullId 
                  (uid 
                    (simpleId CUSTOMERS))))))) WHERE 
        (expression 
          (predicate 
            (predicate 
              (expressionAtom 
                (fullColumnName 
                  (uid 
                    (simpleId CUST_NAME))))) LIKE 
          (predicate 
            (expressionAtom 
              (constant 
                (stringLiteral 'Kash%'))))))))))