I am using H2 to write millions of records into two tables that I then need 
to re-read back out.  The creation of the temp file to store the result set 
is taking about 90 minutes (its a 16gb database) and the utilization is 
really low during this as its just waiting on very slow sequential I/O.  I 
know that server side cursors are on the roadmap and that they are 
complicated.  Are those still "in progress" or are they probably never 
really going to happen?  

So in thinking through my options- I guess I could "partition" the 
resultsets at insert time by some kind of identifier, and then pull out a 
partition at a time such that a single partition was small enough to not 
trigger the temp db creation?  Is there anything else?  A lower-level API 
that I could read each b-tree independantly and hand-roll my own joins with 
lots of prefecthing for the inner join bookmark lookups?  reading will 
happen on a single thread and nothing else will use this database while I'm 
scrolling through these results so its simpler than general server side 
cursors.  Any other options?

Thanks,

Steve

-- 
You received this message because you are subscribed to the Google Groups "H2 
Database" group.
To view this discussion on the web visit 
https://groups.google.com/d/msg/h2-database/-/LtxqfikevqAJ.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/h2-database?hl=en.

Reply via email to