# The use of Druid Sql parser

The use of Druid Sql parser

Import Druid dependencies

<!-- druid -->
<dependency>
    <groupId>com.alibaba</groupId>
    <artifactId>druid</artifactId>
    <version>1.1.23</version>
</dependency>

Application of factory pattern

Java code

• SqlParseService：The interface for specific parsing statements

public interface SqlParseService {
    
    

    /**
     * 解析创建语句
     * @param sql
     * @param dbtype
     * @return
     */
    List<ParseItem> parseCreateSql(String sql,String dbtype);

    /**
     * 解析查询Sql
     * @param sql
     * @param dbtype
     * @return
     */
    List<ParseItem> parseQuerySql(String sql,String dbtype);
}

• SqlParseFactory: Get different parsers according to dbtype: Mysql, Oracle

public class SqlParseFactory {
    
    

    private static final Logger log = Logger.getLogger(String.valueOf(SqlParseFactory.class));

    /**
     * 多个线程见可见
     */
    private static volatile SqlParseFactory sqlParseFactory = null;

    private SqlParseFactory() {
    
    

    }

    /**
     * 线程同步
     * @return
     */
    public static synchronized SqlParseFactory getInstance() {
    
    
        if (Objects.isNull(sqlParseFactory)) {
    
    
            return new SqlParseFactory();
        }
        return sqlParseFactory;
    }

    public SqlParseService getService(String dbType){
    
    
        if(SqlParseConstant.MYSQL.equalsIgnoreCase(dbType)){
    
    
            return new MysqlSqlParse();

        }
        return new OracleSqlParse();
   }
}

• MysqlSqlParse inherits into the abstract class AbstractSqlParse, some common attributes can be encapsulated in the abstract class, and the abstract class implements SqlParseService

public class MysqlSqlParse extends AbstractSqlParse {
    
    

    private static final Logger log = Logger.getLogger(String.valueOf(MysqlSqlParse.class));

    @Override
    public List<ParseItem> parseCreateSql(String sql, String dbtype) {
    
    
        List<ParseItem> list = new LinkedList<>();
        ParseItem parseItem = new ParseItem();
        List<Map<String,String>> column = new LinkedList<>();
        SQLStatementParser statementParser = SQLParserUtils.createSQLStatementParser(sql, SqlParseConstant.MYSQL);
        SQLCreateTableParser sqlCreateTableParser = statementParser.getSQLCreateTableParser();
        SQLCreateTableStatement sqlCreateTableStatement = sqlCreateTableParser.parseCreateTable();
        SQLName name = sqlCreateTableStatement.getName();
        List<SQLTableElement> tableElementList = sqlCreateTableStatement.getTableElementList();
        for (SQLTableElement sqlTableElement : tableElementList) {
    
    
            Map<String,String> map = new HashMap<>();
            if(sqlTableElement instanceof SQLColumnDefinition ){
    
    
                SQLColumnDefinition sqlColumnDefinition = (SQLColumnDefinition) sqlTableElement;
                String columnName = ((SQLColumnDefinition) sqlTableElement).getColumnName();
                SQLDataType dataType = sqlColumnDefinition.getDataType();

                map.put("column",columnName);
                map.put("datatype", String.valueOf(dataType));
            }
            column.add(map);
        }
        parseItem.setDbtype(dbtype);
        parseItem.setColumn(column);
        parseItem.setTable(String.valueOf(name));
        list.add(parseItem);
        return list;
    }

    @Override
    public List<ParseItem> parseQuerySql(String sql, String dbtype) {
    
    
        return super.parseQuerySql(sql, dbtype);
    }
}

• The ASC abstract syntax tree is parsed. There are various APIs in Druid, and you can parse various SQLs and use them directly.