In article <322CA6C8.6D9F@segel.com>, Mike Segel <mikey@segel.com>
writes
>Jonathan Leffler wrote:
>>
>> An alternative is to locate one copy of the job table on one of
>> the servers, and to have the other two databases contain a synonym
>> which points to the remote table. The advantage of this is that
>> it works; the disadvantage is that it relies on the central machine
>> being up.
>>
>> If that is not OK, then replication is probably the answer.
>>
>> I think the performance of a stored procedure implementing the view
>> is very unlikely to be good enough, not least because SPs do not
>> support dynamic SQL so there is no way of restricting the set of
>> rows returned by the view -- the SP would always end up returning
>> all the rows in all three tables.
>>
>> Yours,
>> Jonathan Leffler (johnl@informix.com) #include <disclaimer.h>
>>
>> }From: "Chao Y. Din" <cdin@csc.com>
>> }Date: Mon, 02 Sep 1996 09:29:14 -0400
>> }X-Informix-List-Id: <news.27678>
>> }
>> }David Williams wrote:
>[SNIP]
>> }> We don't want to replicate the data just make all the similar tables
>> }> appear as one table so the PC-base reporting tool (GQL) can be used
>> }> by 'normal' users? i.e. non-techines.
>
>Well, lets look at it another way:
>The poster wants to have three separate sites with no co-mingling
>of the data. Each site is responsible for its local data.
>
>So replication will not work. (Based on his needs.)
>
>-Mikey
>(Let me think about this one ....)
Exactly.
--
David Williams