(1) Yes and (2) No. :-) On 8/30/06, Scott Zhu <[EMAIL PROTECTED]> wrote:
Thanks for the quick response. So you are saying we don't need the remapResults if just the table name is dynamic? As far as for the CDATA, since we need that in some other places (to escape the < >, etc.), we just use it consistently for the statement bodies (anywhere no ibatis keyword is used). That shouldn't hurt right? On 8/30/06, Larry Meadors <[EMAIL PROTECTED]> wrote: > The remapResults attributes should only be required if the result > mapping (i.e., column list) changes. > > In the case you included, it should not be (nor should the CDATA section). > > Larry > > > On 8/30/06, Scott Zhu <[EMAIL PROTECTED]> wrote: > > > > In our application, we need to support dynamic table names, so our statement > > would look something like this: > > > > <select id="getProductIndexByOfferId" remapResults="true" > > parameterClass="java.util.Map" resultMap="productIndexResult"> > > <![CDATA[ > > select p.ProductID > > from $UA_ProductIndex$ p > > where OfferID=#offerID# > > ]]> > > </select> > > > > If I understand it correctly, the remapResults is needed if the selected > > columns or table names can be different at statement execution time (in > > order to regenerate the prepared statement). In our case though, once we > > choose the table name, it will remain the same until we change the > > configuration (happens rarely). Since there's a performance hit by using > > this setting (remapResults), I wonder if it's possible for us not to set it. > > Then when the configuration is changed, is there a way we can force a reload > > of (all the) sqlmap statements on the fly (we don't have an option to bounce > > the app server)? I guess if not, I can still throw away the sqlclient > > instance and re-create it. > > > > Thanks. > > >
