This relates to the question I asked last week about indexing, functions,
etc... on datetime fields
I'm thinking about changing one of the column definitions. Currently it
is a DATETIME and I would like to change it to two columns defined as DATE and
TIME. After a bit of tinkering with ALTER on a test table, it seems to me that
the only way to do that is to dump the table, drop it, recreate the table then
pass it through a PERL script that splits off the two fields.
That the only way?
Alternatively, I could just drop an index on the DATETIME field and use a
lot of between-type queries. That seems a little klunky though.
I'm reluctant to make any changes to theis table as it is rather large and
resides on a slow machine with just a SCSI-2 disk.
Anybody ever face this sort of dilemma?
PGP fingerprint: 063FCE320681C336 78C164FC9B2F91EA