> Well, with the next version having toast, what about parsing the email
> and
> storing the body as TEXT?
Actually, email message should be parsed anyway, at least headers
separated from the body. In most cases it's not enough and one has to store
the headers parsed into separate tuples etc etc.
TOAST is a splendid endeavour, but there were huge archives before, when
we didn't have technology like that at hand. It takes a significant burden
off developer's shoulders, but it can't stop the preservation law. This
burden is placed on your server, both hardware and software parts.
In brief, the strategy here depends on usage pattern. We can't predict
maximal length of message body, so we have to decide beforehand, what parts
should be searchable, indexed and what can just be stored.
As an example, the dirty hack that I call my email archive (it was written
in 40 minutes, it even uses psql to talk to the server), stores message
bodies as BLOBs (it shouldn't, plain files are easier). Searching on
sender's address and subject is enough for me. What's enough for you is up
to you.
--
contaminated fish and microchips
huge supertankers on Arabian trips
oily propaganda from the leaders' lips
all about the future
there's people over here, people over there
everybody's looking for a little more air
crossing all the borders just to take their share
planning for the future
Rainbow, Difficult to Cure