Drew I think you will need to check the objects in memory as well.
Paul On Mar 12, 2011, at 9:14 PM, Drew Thoeni wrote: > I am writing objects to a database in this fashion: > > 1) reading a text file with companies in it. > > 2) fetching a company object based on a unique attribute (domain name) from > the text file using "objectWithFetchSpecificationAndBindings" > > 3) if the object is not found, I'm adding the object using > ec.insertObject(tempCompany) and, after creating 1,000 objects, I invoke > ec.saveChanges(). I'm doing this after 1,000 objects to reduce database > overhead. > > However, I am getting database errors (unique constraint) because there are > duplicate items in the text file and the fetch does not seem to be > discovering these items in the 1,000 as yet unwritten objects. > > I thought a fetch would not only check the database but also check the > editing context to see if the item was there and not yet committed. > > I think I must be handling this incorrectly and there is a pattern that would > accomplish my goal, which is to add companies from the text file to the > database only if the company is not yet in the database, but also do this > efficiently. That is, if I invoke ec.saveChanges() for every record, the > duplicates are discovered but the process is much slower. > > Any suggestions are, of course, greatly appreciated. > > > Drew > > > > > _______________________________________________ > Do not post admin requests to the list. They will be ignored. > Webobjects-dev mailing list ([email protected]) > Help/Unsubscribe/Update your Subscription: > http://lists.apple.com/mailman/options/webobjects-dev/pyu%40mac.com > > This email sent to [email protected] _______________________________________________ Do not post admin requests to the list. They will be ignored. Webobjects-dev mailing list ([email protected]) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/webobjects-dev/archive%40mail-archive.com This email sent to [email protected]
