[Dev] some XML database tests with CMDI records

[Leif-Jöran Olsson]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Den 2012-06-05 10:13, Thomas Zastrow skrev:
> Dear all,
> 
> Indeed, an interesting evaluation which matches our experiences with
> eXist and BaseX. For another project (integrating the TüBa-DZ into
> WebLicht and re-writing of the TigerSearch Query langauge as web
> application), we have choosen to use BaseX. eXist is not really able to
> deal with big amounts of complexe XML data.

I would like to dare you a bit on this broad statement. I can tell you
we have done some work on eXist-db to make known hurdles smaller already
since the tech-preview. For instance the data set loading is 45 minutes
with ft-indexing on Components (alternative given in section Document
root index) with a data set of 438055 cmdi-entries compared to the
stated 15 hours in the wiki page.

To be able to make a proper comparison with the figures in the wiki --
on my machine the tech-preview-2.0 version of eXist-db takes between
80-135 minutes, depending on other machine load, to load all three data
sets with the same ft-indexing mentioned above. So it seems to be a
factor of between 2 to 3 times faster than the tech-preview on ft-indexing.

So mainly it seems to be a configuration and hardware issue.

> But the wrong results from
> BaseX are worrying me a little bit ...
> 
> Scott Martens here in Tübingen is our BaseX specialist (working on the
> project mentioned above), probably he can help with BaseX configurations
> etc.

And I can of course do the same with eXist-db configuration etc.

Cheers,
Leif-Jöran
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.18 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iD8DBQFP1fd+hcIn5aVXOPIRAlsGAKCfomowbDdDJbcnt1naqLe1Z3aCPQCgiYWI
h6ZGDOThsSTxNjxTk6ErjZs=
=LRlJ
-----END PGP SIGNATURE-----