Posted Sat, 27 Sep 2014 02:20:00 GMT by

Reading the information you have about change data capture, I know that RODIN can capture changes from journal images in real time. However, we need to capture changes from many tables. Does each table run as a separate job, or is just one job used?

Also, when we build our data warehouse, we will need to be able to load today's data up to a certain point in time in our nightly batch processing. Any changes after that point are tomorrow's data (price changes due to go into effect the next day for example). How would we do that? We can't just use a set time like 10pm - it has to be a specific point in the job, and could be any time of the night, depending on when it started. Can we send a message to the change data capture job.
Thanks in advance,

PS. RODIN looks like a pretty cool product. Can't wait to try it out.

Jessica

Posted Thu, 16 Jun 2016 13:00:00 GMT by

Hi Jessica,

One Change Data Capture job (we call it a CDC server) can capture changes for up to 300 tables associated with the same journal. If you have different journals, or more than 300 tables to process, you will define as many CDC servers as you need. In fact you can have several servers processing the same journal. This is a good solution for a high volume environment when you have a multi-processor LPAR.

We can Apply the captured data at an EXACT position in a journal. We do this by sending a user defined journal entry to the journal (using a supplied command that you can embed in your nightly job at the required point. Then, when the CDC Server processes that user defined journal entry (which may be within micro-seconds, or maybe a few seconds or minutes later) , an Apply is triggered, which takes all of the captured data and moves it to the Apply tables - becoming today's batch of data. Hopefully that answers your questions, but if you need more information just let me know.

Originally Posted By AJ

You must be signed in to post in this forum.