字符串拆分,单列变成多行,并保留原来的其它字段的数据,准备测试数据:
mysql> create table t1(c1 varchar(32),c2 varchar(32));Query OK, 0 rows affected (0.02 sec)
mysql> INSERT INTO t1 SELECT 'man' c1,'123,234,567,789' c2 UNION ALL SELECT 'guo','-12,-87,-09,132' UNION ALL SELECT 'fang','1024,2387,1209,1032';
Query OK, 3 rows affected (0.02 sec)
Records: 3 Duplicates: 0 Warnings: 0
mysql> select * from t1;
+------+---------------------+
| c1 | c2 |
+------+---------------------+
| man | 123,234,567,789 |
| guo | -12,-87,-09,132 |
| fang | 1024,2387,1209,1032 |
+------+---------------------+
3 rows in set (0.00 sec)
mysql>
需要达到的效果是,左边变成右边,按照逗号拆分出来,单列变成多行记录,而且保留原来的其它列数据,如下图所示:
#,blog原地址:http://blog.csdn.net/mchdba/article/details/53889803,作者mchdba,谢绝转载
这个复杂的实现,单条的sql或者几条sql无法达到,需要临时表存储中间数据,所以编写存储过程,间接实现,存储过程如下:
DELIMITER $$USE `test`$$DROP PROCEDURE IF EXISTS `proc_cur_split_str`$$USE `test`$$CREATE PROCEDURE test.proc_cur_split_str()BEGIN DECLARE v_c1 VARCHAR(20); DECLARE v_c2 VARCHAR(60); DECLARE _done INT DEFAULT 0; DECLARE cur_strs CURSOR FOR SELECT c1,c2 from test.t1 ; DECLARE CONTINUE HANDLER FOR NOT FOUND BEGIN SET _done=1; END; DROP TABLE IF EXISTS test.ZZ_2; CREATE TABLE test.ZZ_2(C1 VARCHAR(20),CN VARCHAR(60)); OPEN cur_strs; FETCH cur_strs INTO v_c1, v_c2; WHILE _done != 1 DO DROP TABLE IF EXISTS test.ZZ_1; CREATE TABLE ZZ_1(CN VARCHAR(60)); SET @b=v_c2; SET @a = CONCAT(CONCAT("insert into test.ZZ_1 values('",REPLACE(@b,',',"'),('")),"')"); -- SELECT @a; PREPARE stmt1 FROM @a; EXECUTE stmt1; -- SELECT v_c1,t.* FROM test.ZZ_1 t; INSERT INTO ZZ_2 SELECT v_c1,t.* FROM test.ZZ_1 t; COMMIT; -- SELECT * FROM test.ZZ_2; FETCH cur_strs INTO v_c1, v_c2; END WHILE; CLOSE cur_strs; SELECT * FROM test.ZZ_2; END;
执行存储过程,得到执行结果,符合心理预期,单列C1变成多行,并且保留了原来的其它字段c1的数据值,如下所示,