T
temays
Guest
Hello Valeriy, Let me try to answer your questions. In regards to privileges on the various sys* tables, Satya was correct in his reply that the privileges are necessary for the initial dataserver setup and for ongoing schema change support. There are several kbase articles that describe why are how the privileges get used.(For example see http://knowledgebase.progress.com/articles/Article/17082?q=sys.argument&l=en_US&c=Product_Group%3AOpenEdge&fs=Search&pn=1 ) In regards to the bulk load process and changing the number of rows sent per Oracle transaction, this is very doable and is really one of the more powerful features of how Pro2 operates. You simply change the "template" file being used for the bulk-load procedure generation and code the scoping to work any way you see fit. Quite sometime ago, I think it was in Pro2 version 3.x, I wrote a special Oracle template that used the send-sql-statement to send a "--Bulk-insert Start" and a "--Bulk-insert End" to serve as a wrapper around the sending of many source side rows. You could even code the template to drop the data into ASCII files and use the Oracle command line bulk loader. This is an extremely fast load solution but in my experience as soon as you drop to ASCII you have to start fiddling with "weirdness" you might see in the data, NULL values, special characters, etc. Let me startup a Pro2Oracle VM and I'll try to knock out a sample bulk load template that will give you some ideas of how to proceed. The thing to remember is that by simply tweaking the default Pro2 templates, an end-user can very quickly put in play some pretty massive modifications to how bulk-load and replication procedures work. It really is one of the feature sets that often gets overlooked. Hope this helps and stay tuned for a sample on how you could code this optional way of bulk loading. Terry
Continue reading...
Continue reading...