Implementing data field splicing in DB2 (LISTAGG() and xml2clob, xmlagg)

1. Use the function LISTAGG()

1.1 Same as Oracle implementation

• For Oracle, you can read the following article:
  Oracle commonly used simple SQL operations (detailed explanations of the usage of various functions such as insert into, merge into, start with connect by prior and regexp_substr) .

1.2 Using LISTAGG() in DB2

1.2.1 About DB2 version

• My version is10.5,After testing, DB2 V9.7.0.10 can also be used (I don’t know about other versions, it seems to be 9.7 and above), if you don’t know your own version, you can use the following sql to check:

  select * from sysibmadm.env_inst_info;
  

  

1.2.2 Data preparation

• For data preparation, you can read the following article. It is just the Oracle version, but it is similar and has almost no changes, as follows:
  oracle recursive query (start with connect by prior) And tree statistics connect_by_root (child nodes are summarized to parent nodes).

1.2.3 Code implementation

• Look at the data in the table SYS_USER first, as follows:
  
• Splice the names of all developers under 部门D001 into one field, which is actually exactly the same as Oracle, as follows:

  SELECT
  	DEPT_ID ,
  	listagg(su.USER_NAME  ,',') WITHIN GROUP (ORDER BY su.USER_NAME) all_user_names
  FROM
  	SYS_USER su
  WHERE
  	DEPT_ID = 'D001'
  GROUP BY
  	DEPT_ID ;
  

  

2 Solve the problem of excessive length of LISTAGG() in DB2

2.1 Using xmlagg + xmlelement

• xmlaggThe syntax is as follows:

  xmlagg(xmlelement(NAME [自定义xml标签], '需要拼接的数据')) AS [字段别名]
  

• xmlelement,xmlelement() is a function that specifies XML elements. This function requires the following two parameters:
  • name: Specifies the name of the generated XML tag element.
  • [Data that needs to be spliced]: The data contained in the element can be one or more values, which can be spliced ​​and separated by custom delimiters.
  • xmlagg()The function will concatenate the results of all XML elements into an XML document in a sequential manner, thereby returning an XML type value.
• Implementation code:

  SELECT DEPT_ID ,xmlagg(xmlelement(NAME userName, su.USER_NAME||',')) AS all_user_names
  FROM SYS_USER su 
  WHERE DEPT_ID = 'D001'
  GROUP BY DEPT_ID;　
  

  
• The implementation code (plus sorting) is as follows:

  SELECT DEPT_ID ,xmlagg(
  	xmlelement(NAME userName, su.USER_NAME||',') ORDER BY su.USER_NAME
  ) AS all_user_names
  FROM SYS_USER su 
  WHERE DEPT_ID = 'D001'
  GROUP BY DEPT_ID;　
  

2.2 Remove the xml tag

• Use replace() function, as follows:

  SELECT [分组的字段名]
  ,replace(replace(xml2clob(xmlagg(xmlelement(NAME a, [需要拼接的字段名]||','))),'<A>',''),'</A>',' ') 
  FROM [表名] GROUP BY [分组的字段名];
  

• The test code implementation is as follows:

  SELECT DEPT_ID ,
  replace(replace(xml2clob(xmlagg(xmlelement(NAME userName, su.USER_NAME||','))),'<USERNAME>',''),'</USERNAME>',' ') AS all_user_names
  FROM SYS_USER su 
  WHERE DEPT_ID = 'D001'
  GROUP BY DEPT_ID;