SQL语句使用模式删除字符串的一部分

Question

I have a table column with semicolon delimited values illustrated as follows:

table1.ID | table1.column1 
-----------------------------------------------------
   1      | test1=aaa;test2=bbb;test3=ccc;
   2      | test1=akkaa;test2=bkjkjb;test3=cckjklc;
   3      | test1=ajaa;test2=bjhjbb;test3=sss;

I need to remove all test2=xxxx where xxxx can be of any value.

I started the following query which I need to complete; how I can specify a given pattern / expression?

UPDATE table1
SET column1 = Replace(column1, 'test2=', '')
WHERE column1  Is Not Null;

Thanks in advance

Answer 1

For MSSQL:

Try this query: (I didn't used any pattern[I don't know the possibility]. What I did is replace the string content of test2=something; with empty string. The CharIndex function is used to locate the Test2 and test3 location.)

 UPDATE table1 SET Column1  = 
      REPLACE(Column1, substring(Column1, CharIndex('test2=', Column1), 
           (CharIndex( 'test3=', Column1) - CharIndex('test2=', Column1))), '')

--Live Demo Here

Answer 2

for oracle:

update table1
set column1=substr(table1.column1,1,instr(table1.column1,';',1,1)-1)||substr(table1.column1,instr(table1.column1,';',1,2));
commit;