iBATIS does not cache results unless you specifically ask it to do so. And there is no "phantom table" code in iBATIS (this did make me smile when I read it). Have you configured any <cacheModel> elements anywhere? If not, then iBATIS is not caching results. If you have configured <cacheModels>, then temporarily disable them to see if anything changes.
More likely this is something related to transaction configuration. Or maybe explicit transactions are not being committed/ended properly somewhere.
Jeff Butler
On 8/1/06, Albert L. Sapp <[EMAIL PROTECTED]> wrote:
I have run into the same thing where anything done outside of the
application (isql*plus or enterprise manager) does not appear in the
results from a query inside iBatis. We found if we restarted the tomcat
server that the activity done outside the application was now visible to
the application. We stopped modifying table data directly and all
interaction is now through the application. If we absolutely have to
work outside the application, we know we must do a restart of the tomcat
server.
It has been a while since we had to deal with this so we never really
tried to find out why this happens. Something to do with data caching,
I am sure. We know that we will have the problem, so we just deal with
it the way we know works.
Al
cmose wrote:
> Hey all,
> I'm mostly an ibatis noob and having quite a bizarre problem, for me at
> least, with ibatis returning phantom results on some test queries.
>
> Basically, is there any caching happening with Ibatis behind the scenes that
> I should be aware of?? Right now I'm using a simple data source and letting
> ibatis handle the conneciton pooling and etc., I seem to be having a
> situation where queries in the ibatis mapping file return results different
> from those same queries pasted into Oracle's isql*plus (e.g., the queries
> run from ibatis only return results if they were inserted using ibatis).
>
> Ibatis seems to be using a phantom table when it's running it's inserts and
> queries, that completely ignores anything that has been done to the table
> via isql*plus (e.g., pasting the sql code from the mapping file directly
> into isql*plus yields vastly different results then running it using
> ibatis).
>
> Now I realize that something could be wonky with isql*plus, but I'm more
> likely to believe a configuration or use error on my part then isql*plus not
> working correctly.
> Thanks much!!
>
> As an aside to this, is there anything in particular, configuration wise or
> functionality wise that I sould worry about when using a singleton
> SqlMapClient across a dozen or so classes?
>
> (sorry if this was a repeat, I received a nabble confirmation in my email
> and made the mistake of assuming that was also the list subscription
> confirmation for which I received another email after already having posted
> this, ergo I'm uncertain whether or not the original message actually went
> out - thanks for your patience ;-) ).
>
