Thanks Thomas, this is very useful to know.

We are using H2 in a number of places very happily. One of these is
for a key/value store.  The original plan was that the H2 key/value db
would be just a short-term stub and would be replaced with our own
implementation or an existing key/value style db.  To get the H2
version to perform we sharded the db to reduce lock contention,
queueing up merges so that they could be done in bulk, and so on.  By
the time we got around to experimenting with key/value-style dbs such
as Berkeley we were surprised to discover that H2 implementation was
significantly faster.  It's also been extremely robust.

Recently however, the size of the data has grown large enough that the
H2 implementation doesn't seem to be scaling well.  I haven't had a
chance to see exactly why yet but I suspect that at least part of the
problem is that the whole b-tree index is now longer staying in
memory.   We will probably end up using Cassandra for the key/value db
and happily continue to use H2 in other places.   I know that this
particular use is outside the range that H2 is targetted for, but it
does seem to me that the range and scale of H2 might be usefully
extended if it were possible to create in-memory hash indexes for
persistent tables.

Regards,

Paul

On Mar 20, 4:03 am, Thomas Mueller <[email protected]>
wrote:
> Hi,
>
> The documentation forhashindexes is incorrect, I will try to fix it.
> Here is the new documentation:
>
> ---------------Hashindexes are meant for in-memory databases and memory tables
> (CREATE MEMORY TABLE). For other tables, or if theindexcontains
> multiple columns, theHASHkeyword is ignored.Hashindexes can only
> test for equality, and do not support range queries (similar to ahash
> table). Non-unique keys are supported.
> ---------------
>
> Regards,
> Thomas

-- 
You received this message because you are subscribed to the Google Groups "H2 
Database" group.
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