Ramesh, (keep discussion on the list for all)
> 1) May it is because that in BMP you have solid control over the sql and it > is very easy to debug than CMP? You shouldn't have to debug CMP SQL, so imho no. > 2) Also may be because BMP require for a collection n-EB's n+1 DB calls to > laod the data? Where as in CMP it is simply one call with huge SQL Most app servers these days avoid the n+1 problem with read-ahead mechanisms, so again, imho no. > 3) do you know about fat key pattern to increase the BMP performance? I am aware of it, but probably dont have intimate enough knowledge to comment intelligently. At a high level, is there any reason why a CMP implementation could not also implement this? > 4) When you say that in BMP you have greater control But actually when you > change the DB from say DB2 to Oracle there are some things which doesnot > like. In that case you need to change the whole SQL code test and compile > and re-run and deploy. That's a huge chunk. Whereas in CMP this is not at > all there. When porting from db to db there is always the risk - regardless of where the db interaction is handled. I have only recently found out that apparently DB2/AS400 returns odd values for null primitives. These sort of problems will have to be handled at some level, its just a matter of whether your container handles it or you. Sometimes when a container handles it you still have to do stuff. cheers dim > > > ??Ramesh Kesavanarayanan > ? [EMAIL PROTECTED] > > > =========================================================================== To unsubscribe, send email to [EMAIL PROTECTED] and include in the body of the message "signoff EJB-INTEREST". For general help, send email to [EMAIL PROTECTED] and include in the body of the message "help".
