Data Warehouse:

My current job has me working with data warehousing a lot more than I have up to this point. There’s some really cool stuff going on in this area. I’ve been tearing through a couple good books on the ‘Kimball’ method and this week I’ve viewed several hours of webcast from Microsoft.

There seam to be two competing philosophies around data warehousing. I’ve been reading and practicing the ‘Kimball’ method, named after Ralph Kimball. It’s interesting stuff. It’s not at all like the transactional systems I’ve been supporting up to this point. In a nutshell, you do a lot of the work of summarizing the data during the data load. When it’s time to make reports you just pull the summarized data you pre-calculated, or you aggregate new summaries for whatever you need. For example, you load all your data into tables, but then load other tables with aggregate values (like total sales per day). Later you can pull the daily totals, or you can sum the daily totals over a month, much faster than if you had to calculate the totals for all the detail records. Now that’s cool.