CODEWITHSUNDEEP
sqlselectpivotteradata
Petr Novák
Petr Novák

Reputation: 115

sql rows into columns pivot table

I have sql view like this:

id date_from  date_to
1  2005-01-05 2005-05-10
1  2005-01-05 2005-05-10
1  2005-01-05 2005-05-10
2  2005-01-01 2006-08-31
2  2010-01-01 2010-06-30
3  2005-01-01 2005-06-30

And I want to write sql statement which returns:

1 2005-01-05 2005-05-10 2005-01-05 2005-05-10 2005-01-05 2005-05-10
2 2005-01-01 2006-08-31 2010-01-01 2010-06-30 NULL       NULL
3 2005-01-01 2005-06-30 NULL       NULL       NULL       NULL

Any ideas?

Upvotes: 1

Views: 2034

Answers (2)

Osa E
Osa E

Reputation: 1781

Answers to Mark's question will help.

Which RDBMS is this - MySQL, Oracle, SQLServer...? Also, are you ever going to want more than three date_from, date_to pairs per id?

using t-sql, I explicitly handle 3 levels. If you want it to be dynamic, you need to create the query dynamically.

DECLARE @staging TABLE
(
    id int NOT NULL,
    date_from datetime NOT NULL,
    date_to datetime NOT NULL,
    step int
)

INSERT INTO @staging
SELECT id, date_from, date_to,
    ROW_NUMBER() OVER(PARTITION BY id ORDER BY date_from, date_to)
FROM tblTemp

-- below is static for 3 levels, make below dynamic to match what you want
SELECT t1.id, t1.date_from, t1.date_to, t2.date_from, t2.date_to, t3.date_from, t3.date_to
FROM @staging t1 LEFT OUTER JOIN
    @staging t2 ON t1.id = t2.id AND t2.step = 2 LEFT OUTER JOIN
    @staging t3 ON t2.id = t3.id AND t3.step = 3
WHERE t1.step = 1

Test @ http://sqlfiddle.com/#!3/9daae/4/0

Upvotes: 1

sgeddes
sgeddes

Reputation: 62831

Because you're most recent questions have been related to MySQL, I'll assume you want a MySQL solution.

If you know the max number of potential date ranges, then you can use MAX and CASE. However, you have to have a row counter since you don't have any other unique identifier (I'd actually recommend adding that to your view since you mention this is a view). Here it is though:

SELECT Id,
   MAX(CASE WHEN row_number = 1 THEN date_from END) date_from1,
   MAX(CASE WHEN row_number = 1 THEN date_to END) date_to1,
   MAX(CASE WHEN row_number = 2 THEN date_from END) date_from2,
   MAX(CASE WHEN row_number = 2 THEN date_to END) date_to2,
   MAX(CASE WHEN row_number = 3 THEN date_from END) date_from3,
   MAX(CASE WHEN row_number = 3 THEN date_to END) date_to3
FROM (
  SELECT 
      id, 
      @running:=if(@previous=id,@running,0) + 1 as row_number,
      @previous:=id,
      date_from, date_to
  FROM YourView 
      JOIN    (SELECT @previous := 0) r
    ORDER BY id, date_from, date_to
) t
GROUP BY Id

If you don't know the maximum number of date ranges, then you won't be able to do this with a single SQL statement. Instead, you'll need to use dynamic SQL.

I'll assume you can add the row_number to your View and here is a close example:

SET @sql = NULL;
SELECT
  GROUP_CONCAT(DISTINCT
    CONCAT(
      'MAX(CASE WHEN row_number = ', 
      row_number, 
      ' THEN date_from END) date_from', 
      row_number)
  ),
  GROUP_CONCAT(DISTINCT
    CONCAT(
      'MAX(CASE WHEN row_number = ',  
      row_number, 
      ' THEN date_to END) date_to',  
      row_number)
  ) INTO @sql
FROM YourView;

SET @sql = CONCAT('SELECT  ID, ', 
                   @sql, ' 
                   FROM    YourView
                   GROUP   BY ID');

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

Best of luck.

Upvotes: 1

Related Questions