If I have large articles that need to be stored in a database, each associated with many tables would a NoSQL option help? Should I copy the 1000 char articles over multiple "buckets", duplicating them each time they are related to a bucket or should I use a normalized MySQL DB with lots of Memcache?
+2
A:
When storing data normalising it is essential, and will save you time further down the line 99% of the time, even if you think you are saving time at the time. Always normalise!
Also you have terminologoy problem, de-normilisation refers to taking normalised data and breaking it down and restoring it as repeated data.
Tom Gullen
2010-06-16 13:39:37
Agree. The only exception to normalization is if you have HUGE tables that you need to query against a lot (and hence a lot of joins). Then, you can gain quite a bit by denormalization. But personally I always keep a fully normalized copy along side a denormalized copy. Some querieis are FAR better with a normalized set, and others are better with a denormalized set. Maintaining both increases your flexibility (At the price of disk space and complexity)...
ircmaxell
2010-06-16 13:44:04