Posts Tagged ‘data warehouse’


The purpose of Merge is where you have two tables – maybe a transaction table and a staging table which could have records to be inserted and/or replaced in the transaction table.  Or maybe you have two tables of people (staff or patients) – one table (the master) with last years data and one table from this year, and you want to bring the master up to date with additions, updates or removals. Under SQL 2005 you would have had to


There is a bug in SQL Server 2008 and 2005.  When you try and link a 64 bit server to a 32 bit server, the select of data can fail.   But the fail is not obvious. When you say "select * from linkedserver.mydbname.dbo.table" it will return a message about stored procedures such as: The stored procedure required to complete this operation could not be found on the server. Please contact your system administrator. This is because


Things to check when your data warehouse loads take too long‏

If you ever find yourself suffering from slow data load problem, let me say without a doubt…. there is no situation where the loading of data into a data warehouse should go beyond a couple of hours. If a data warehouse here at WhippsX NHS Trust can load 2 years worth of data from 3 different trusts in less than 2 hours, any data warehouse should be able to load a days worth of data in a few


Whenever you INSERT data into SQL tables, as each row is added, it also updates the index data on a row by row basis.  So a table with 15 indexes will actually update 16 rows (1 data and 15 index rows) for each insert. Where you are inserting thousands of rows, this can make for MASSIVE log files and LONG processing times. So, to make insertion super fast, always a good idea to DISABLE non-clustered indexes (those indexes which