You're right, this isn't a Tapestry-specific problem. I doubt there is a solution that doesn't involve compromises, but there are some interesting suggestions here: http://stackoverflow.com/questions/25099732/paging-of-frequently-changing-data
On Nov 30, 2016 3:38 AM, "Adam X" <vbgnm3c...@gmail.com> wrote: Howdy again! This question is probably not so much about Tapestry but from what I see on this list some of you have build and/or are maintaining large enterprise apps in T5. I'm building such an app and I'm facing a dillema on a TransactionView page (navigated from TransactionSearch page). I'm not asking about how to build paged grid because jumpstart has very good examples and I'm good with that. My problem is that my transaction data source is extremaly active with many transactions coming in real time. When one searcher for transactions and gets N results which are then paged, say 1-10 of N, and then navigates to page 2 expecting to retrieve 11-20 of N but in reality it will be 11-20 of N+K where K is the number of transaction that has arrived after rendering page 1 before requesting page 2. So 11-20 will repeat K transactions from page 1 because of K number of new transactions (which would now be on page 1). I can think of solving it in 2 ways, neither of which I want to do for obvious (performance) reasons: 1) Fetch all records and persist them (ugh!) 2) Create a temporary table. Option 2 seems just as bad and actually only works for RDBMS data source. Unfortunately once of my sources is DynamoDB so I even wouldn't be able to use option #2 if I wanted to. How do you guys solve this kind of problem? Adam --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tapestry.apache.org For additional commands, e-mail: users-h...@tapestry.apache.org
