PIVOT operator in SQL2005 helps us to traversing rows to columns. For that we would need to specify the row names that we want to see as colums in the columns list. And if a column name, that does not exist as a row is specified, it shows NULL against those columns. But i found this a bit inconsistent in a specific scenario. This is with Beta 2
For example let me assume row 'CYCLE' does not exist and during PIVOTing, i include it in the column list, as shown below.
SELECT * FROM SalesOrder PIVOT (SUM(Quantity) FOR Product IN ([Bike],[Chain][Cycle])) AS PVT
Resutset is like:
Customer Bike Chain Cycle
1 2 3 NULL
2 2 5 NULL
The above resultset is expected and correct.
But if i give a name similar to the one of the other colums, say 'Chain1' which also does not exist, then the resultset is wrong. It's carrying the values of Chain to Chain1 as well.
SELECT * FROM SalesOrder PIVOT (SUM(Quantity) FOR Product IN ([Bike],[Chain][Chain1])) AS PVT
Resutset is like:
Customer Bike Chain Chain1
1 2 3 3
2 2 5 5
Seems like it's finding the best match :)
Could come back to blogging after long time!!
So many interesting things have been happening in the past 2 month....
Starting with Satyam Architect Conference, TechEd 2005, me being MVPd :), got certified in PSP-TSP and several other cool things...
But i also experienced the frustration due to low internet bandwidth leading to my low online presence. Had to miss several MSDN chat sessions, live meetings, forum activities,blogging and also friends online because of this :(
But never mind; MUGH always helps me to be busy offline too :)
MUGH is planning for series of VS2005 and SQLServer contests and sessions.
Part of this is the Whidbey Coding Contest and SQLserver Whitepaper contest.
There is also a plan for having series of Whidbey sessions called 'Deep Impact'.
Complete information on this would be posted soon on www.mugh.net