From: Steve S. <sap...@gs...> - 2003-05-10 11:54:29
|
I brought this topic up a while ago here. Chris gave me some starting pointers which I sat on while I caught up on my "real" work. I recently had some time to get back to this. Here's my current dilemma: We have a package that we use for data transforms that allows you to configure a set of inputs/filters/outputs. Basically you build a map defining all the steps. Once you have a map you can either feed it to a translator, which runs the whole thing; or you can fetch a row from an output node and do what you want with that row. There are no row level IDs used here -- if you're fetching, you just fetch the next row until there are no more. SPOPS, being more of a table model, wants IDs for fetches. Is there a way around this? Looking at what we have, it's really more suitable as a template level tool maybe; i.e., just give it a wrapper and expand the data sets using TT constructs. But I'd like to utilize the security layer SPOPS offers for data access. And, since we do use this for canned data writes/updates, I want to keep an eye on plugging this in more fully as a full data access layer. In all cases where I'd be using this now (read only), I'd be fetching the entire set of data the translation defines, although I will want to be able to segment larger sets so they can be paged. Just using implicit sequence numbers seems to fit. It's not clear to me how far up the need for an ID bubbles though ... assigning my own would not give higher level users any universally known key (like a user name, etc.). What sort of usage impact does that have within the entire OI framework? Also, regarding the paging need, I'd want to not be required to load the entire set in since some of these can be quite large. Load as needed is mostly the model I'd want. -- Steve Sapovits GSI Commerce, Inc. http://www.gsicommerce.com Email: sap...@gs... |