Why not have a seperate table that will randomize for you isntead of associating random data elements in your primary search criteria?
Teddy On 9/28/06, James Smith <[EMAIL PROTECTED]> wrote: > > I am trying to select rows from a database that are both random AND > distinct... > > Simple no? Well... No. > > I can select distinct records using > > SELECT DISTINCT TOP 5 t1.Field1, t2.Field2 > FROM t1 JOIN t2 ON (some id's) > > Now normally to get random records you would just order by newid()... > > SELECT DISTINCT TOP 5 t1.Field1, t2.Field2 > FROM t1 JOIN t2 ON (some id's) > ORDER BY newID() > > But when the DISTINCT keyword is used the "order by" fields must be part > of > the select... > > SELECT DISTINCT TOP 5 t1.Field1, t2.Field2, newID() > FROM t1 JOIN t2 ON (some id's) > ORDER BY newID() > > Unfortunately, this makes all records unique as they all now have a random > field in them! > > How can I randomise these records while having them remain distinct? > > -- > Jay > > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| Introducing the Fusion Authority Quarterly Update. 80 pages of hard-hitting, up-to-date ColdFusion information by your peers, delivered to your door four times a year. http://www.fusionauthority.com/quarterly Archive: http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:254595 Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
