[13:11:13 CDT(-0500)] <cwaymire> EricDalquist: you around? got a hibernate question you may be able to assist with.
[13:11:54 CDT(-0500)] <EricDalquist> sure
[13:13:30 CDT(-0500)] <cwaymire> so I have an xml document that represents an Topic, including all of the Announcement children and am using JAXB to unmarshal that into a Topic object. Passing that to the appropriate service to persist it, everything gets persisted properly except that the key column that binds the Announcement to the Topic (PARENT_ID) remains null.
[13:13:42 CDT(-0500)] <EricDalquist> hrm
[13:13:52 CDT(-0500)] <cwaymire> I thought that being as it is the key column for the mapped set, it would take care of it
[13:13:55 CDT(-0500)] <EricDalquist> so as a rule I always have had a seperate XML vs DB data model
[13:14:14 CDT(-0500)] <EricDalquist> just because the data models never match up exactly
[13:14:32 CDT(-0500)] <EricDalquist> how are you describing the Topic object to hibernate?
[13:14:33 CDT(-0500)] <cwaymire> I was trying to use some simply jaxb annotations to be able to export/import easily
[13:14:39 CDT(-0500)] <EricDalquist> yeah
[13:14:42 CDT(-0500)] <EricDalquist> jen tried that too
[13:14:48 CDT(-0500)] <cwaymire> it currently using a hibernate mapping xml document
[13:14:49 CDT(-0500)] <EricDalquist> it never quite worked well
[13:15:06 CDT(-0500)] <EricDalquist> we're actually just finishing a jpa2 retrofit on a portlet
[13:15:11 CDT(-0500)] <cwaymire> i could have it export the ID values but thats not something i'd like to do
[13:15:19 CDT(-0500)] <cwaymire> bad practice
[13:15:21 CDT(-0500)] <EricDalquist> and will be publishing documentation and some utility libraries in about a week
[13:15:21 CDT(-0500)] <EricDalquist> right
[13:15:36 CDT(-0500)] <EricDalquist> which is why uportal and lots of other things have seperate data models
[13:15:46 CDT(-0500)] <EricDalquist> the export data model is close but not exactly the same as the DB data model
[13:16:02 CDT(-0500)] <EricDalquist> and you go insane trying to get all the little id/reference cases working
[13:16:13 CDT(-0500)] <cwaymire> ok. i'm going to poke at it a little bit more before moving on 
[13:16:18 CDT(-0500)] <EricDalquist>
[13:16:34 CDT(-0500)] <EricDalquist> I'll be sure to send some emails when we get our jpa docs posted
[13:16:43 CDT(-0500)] <cwaymire> cool. thanks.
[14:37:31 CDT(-0500)] <cwaymire_> EricDalquist: as an update. I found why it wasn't updating the child's parent key column.
[14:37:38 CDT(-0500)] <EricDalquist> oh?
[14:37:40 CDT(-0500)] <EricDalquist> what was going on
[14:38:02 CDT(-0500)] <cwaymire_> the definition of the set on the parent was flagged as inverse=true so the child owned the relationship
[14:38:25 CDT(-0500)] <EricDalquist> ah
[14:38:28 CDT(-0500)] <EricDalquist> yeah that would do it
[15:12:30 CDT(-0500)] <cwaymire_> able to successfully export the announcements data to xml and re-import it
[15:12:41 CDT(-0500)] <cwaymire_> using the maven lifecycle hook
[15:13:44 CDT(-0500)] <jwennmacher> good job.
[15:49:13 CDT(-0500)] <cwaymire_> isn't the Concurrent user aggregator supposed to aggregate on unique logins?
[15:49:36 CDT(-0500)] <EricDalquist> no
[15:49:40 CDT(-0500)] <EricDalquist> it watches all events
[15:49:44 CDT(-0500)] <cwaymire_> ahhh
[15:49:57 CDT(-0500)] <EricDalquist> and tracks "active" users by those it has seen an event from in that interval
[15:50:05 CDT(-0500)] <EricDalquist> and it isn't actually by user
[15:50:08 CDT(-0500)] <EricDalquist> it is by event session id
[15:50:13 CDT(-0500)] <cwaymire_> ahh ok
[15:50:22 CDT(-0500)] <EricDalquist> so if you login once
[15:50:31 CDT(-0500)] <EricDalquist> but stay active in the portal clicking around
[15:50:36 CDT(-0500)] <EricDalquist> you will be "active" for a long time
[15:52:13 CDT(-0500)] <cwaymire_> its been requested to modify the Portal Activity portlet to take into account unique logins for the totals so i'll have to modify it a bit.
[15:52:41 CDT(-0500)] <EricDalquist> yeah concurrent user aggregator won't get you that data
[15:52:48 CDT(-0500)] <EricDalquist> you'd have to look at unique logins
[15:52:57 CDT(-0500)] <cwaymire_> yep