如何使用MySQL作为动态数据透视表或交叉表

Question

I know this question or a variance thereof has been asked multiple times, and I have even tried implementing a solution, but I am struggling in completing it.

I have a very simple table with three data columns: Date1, Report_#, Name. I would like to pivot it around Date and the various names as the header tab, and have all the report numbers appear below.

So it would LOOK like this:

Report | Date   | Name                     Date   | Name1 | Name2 | Name3
-----------------------                    ------------------------------
1      | 4-5-12 | Name1                    4-5-12 | 1     | 2     | 3
2      | 4-5-12 | Name2      ----->        4-6-12 | 4     | 5     | 6
3      | 4-5-12 | Name3                    4-7-12 | 7     | 8     | 9
4      | 4-6-12 | Name1                    
5      | 4-6-12 | Name2                    
6      | 4-6-12 | Name3                    
7      | 4-7-12 | Name1
8      | 4-7-12 | Name2
9      | 4-7-12 | Name3

I was able to get an idea of what to do from http://www.artfulsoftware.com/infotree/queries.php#78 but I'm stuck.

I was able to isolate the column names manually and list the report # under each name, but I want to do dynamically figure out the distinct names, make them column names, and list the corresponding reports.

So I created a procedure that would find distinct names and output the correct code. Now I have a hard time plugging the results of the procedure onto the query. And the link above does not help whatsoever, (it seems to skip that step).

So here is the code for the manual way:

SELECT `DATE`, 
GROUP_CONCAT(CASE `Name` WHEN 'Name1' THEN `Report` END) AS 'Name1', 
GROUP_CONCAT(CASE `Name` WHEN 'Name2' THEN `Report` END) AS 'Name2' 
FROM `report_db` GROUP BY `DATE` ORDER BY `DATE`;

And here is the code that dynamic prints the GROUP_CONCAT(... for all the distinct names in the database:

DROP PROCEDURE IF EXISTS writecountpivot;
DELIMITER | 
CREATE PROCEDURE writecountpivot( db CHAR(64), tbl CHAR(64), col CHAR(64) ) 
BEGIN 
DECLARE datadelim CHAR(1) DEFAULT '"'; 
DECLARE singlequote CHAR(1) DEFAULT CHAR(39); 
DECLARE comma CHAR(1) DEFAULT ','; 
SET @sqlmode = (SELECT @@sql_mode); 
SET @@sql_mode=''; 
  SET @sql = CONCAT( 'SELECT DISTINCT CONCAT(', singlequote, 
                     ',group_concat(IF(', col, ' = ', datadelim, singlequote, comma, 
                     col, comma, singlequote, datadelim, comma, '`IR NO`,null)) AS `',  
                     singlequote, comma, col, comma, singlequote, '`', singlequote,  
                     ') AS countpivotarg FROM ', db, '.', tbl, 
                     ' WHERE ', col, ' IS NOT NULL' ); 
  -- UNCOMMENT TO SEE THE MIDLEVEL CODE: 
  -- SELECT @sql;  
  PREPARE stmt FROM @sql; 
  EXECUTE stmt; 
  DROP PREPARE stmt; 
  SET @@sql_mode=@sqlmode; 
END
| 
DELIMITER ; 
CALL writecountpivot('database','`report_db`','name');

and the result of the code above would be this

,group_concat(IF(name = "Name1",`IR NO`,null)) AS `Name1`
,group_concat(IF(name = "Name2",`IR NO`,null)) AS `Name2`
,group_concat(IF(name = "Name3",`IR NO`,null)) AS `Name3`

** So how do I take this text and plug it to my SQL? How do I put the procedure and the query together?**

Answer 1

It's difficult to tell exactly what it is you're trying to do, but if you're attempting to create a procedure that can perform arbitrary pivots then you'll need to provide it with more arguments (such as the column on which to pivot and the column in which values can be found).

Furthermore, you will need to create a prepared statement from within a prepared statement. The outermost statement will use GROUP_CONCAT() to construct the GROUP_CONCAT() expressions that are to be executed, based on the unique values in the specified column:

CREATE FUNCTION SQL_ESC(_identifier VARCHAR(64))
RETURNS VARCHAR(130) DETERMINISTIC
RETURN CONCAT('`',REPLACE(_identifier,'`','``'),'`')//

CREATE PROCEDURE writecountpivot(
  IN _db_nm VARCHAR(64),
  IN _tb_nm VARCHAR(64),
  IN _cl_gp VARCHAR(64),
  IN _cl_pv VARCHAR(64),
  IN _cl_vl VARCHAR(64)
) BEGIN
  SET @sql := CONCAT(
    "SELECT CONCAT('
              SELECT   ",SQL_ESC(_cl_gp),",',
              GROUP_CONCAT(DISTINCT CONCAT(
                'GROUP_CONCAT(IF(",
                   SQL_ESC(_cl_pv),"=',QUOTE(",SQL_ESC(_cl_pv),"),'
                 , ",SQL_ESC(_cl_vl),"
                 , NULL
                 )) AS ',SQL_ESC(",SQL_ESC(_cl_pv),")
              )), '
              FROM     ",SQL_ESC(_db_nm),".",SQL_ESC(_tb_nm),"
              GROUP BY ",SQL_ESC(_cl_gp),"
            ')
     INTO   @sql
     FROM   ",SQL_ESC(_db_nm),".",SQL_ESC(_tb_nm)
  );

  PREPARE stmt FROM @sql; 
  EXECUTE stmt; 
  DEALLOCATE PREPARE stmt;

  PREPARE stmt FROM @sql; 
  EXECUTE stmt; 
  DEALLOCATE PREPARE stmt;
END//

See it on sqlfiddle .