On Sun, Jun 24, 2007 at 06:45:00PM -0000, danmcb wrote: > into two tables, one for the originals, one for the translations (the > objects are actually phrases in many languages). But I'd rather avoid > that because in all other ways, the objects have the same properties, > reference the same objects, and so on. Splitting into two types is not > really desirable, from the point of view of what they are modelling. This sounds to me like you have a one:many possibility, and you should split your columns up instead. (In general, I find, when you have a self-join, it's probably an indicator that some of the data is repeated, and that tells you you're not normalised correctly.) A -- Andrew Sullivan | ajs@xxxxxxxxxxxxxxx The plural of anecdote is not data. --Roger Brinner
