BoxerFlat4

One for the MySQL geeks out there.....

I'm currently planning out some database's, and I've automatically designed them with the Primary Key to be INT's, so I would have..

UID = 0
UID = 1
... etc etc.

However, would I suffer any penalty for designing the Primary Key to be, say a string ? That way, I could have instead....

UID=HKS.Turbo.Thumbnail
UID=Blitz.Turbo.ThumbNail
... etc etc.

Which would be far more easier to read. Are there any issue's with doing it this way ?

GaryK

Boxer,

You will normally find databases handle integer indexes much faster than string ones. If you are using a string you would need to a create a UNIQUE index which would probably affect performance also.

Besides I always follow the rule of never using business data to form a primary key (there are some who oppose this!) but it basically means if you need to change say a part no. you dont need to re-code the db as any relationships are defined on a separate integer primary key.

HTH


Gary


[Edited by GaryK - 11/27/2002 8:34:21 AM]

BoxerFlat4

Cheers Gary -

So the only real objections are design ones : it should in theory work ? We're talking about a database of approx. 5,000 items at the moment, so it's not a monster, but I have normalized it so I've actually got around 8 seperate tables, each which will have 5,000 items.

I don't think performance would suffer too much, but I want to get it right in the design stage first !

GaryK

Boxer,

Yes in theory it would work, you just need to make sure you have a unique index constraint to prevent duplicates.

Gary