« Sargability, hash joins, and the Query Store .... | Main | More about UNIQUEIDENTIFIER/GUID columns in clustered keys .... »

11/15/2018

Comments

Howard Van Dam

Nice column! I had never heard of the term "Sargability" and had to look it up. As a young developer at IBM, we were told using LIKE, and other non SARGable operations was not only bad form, but pointed to an error in the data architecture. At that time we had people dedicated to properly analyzing the data and designing the database. I believe that today we are sorely lacking in the skills these data architects possessed. My last job used a mySQL database that was the most pathetic mess I had ever seen. The database was "developed", and I use that term very loosely, by the originator of the product. Very poor design, and used exactly what you recommend against, the use of a GUID value as both a primary key and a column in a clustered index. It had horrible performance, was impossible to maintain and extend as the product evolved.

The comments to this entry are closed.