Dear all,
when testing BaseX with the TPoX benchmark workload (http://tpox.sourceforge.net/) I could not load the input data set since it contains elements with more than 32 attributes.
The large number of attributes stems from the FIXML data (http://www.fixprotocol.org/specifications/fix4.4fixml) contained in the workload, so it is not just a stress test by the benchmark designers.
Is there any way to lift this limitation?
Best Regards, Peter
Hi Peter,
currently the number of attributes for an element is limited to 32. Maybe we will increase the number of attributes for an element in a future release.
Kind regards, Andreas
Am 13.12.10 11:50, schrieb Peter Fischer:
Dear all,
when testing BaseX with the TPoX benchmark workload (http://tpox.sourceforge.net/) I could not load the input data set since it contains elements with more than 32 attributes.
The large number of attributes stems from the FIXML data (http://www.fixprotocol.org/specifications/fix4.4fixml) contained in the workload, so it is not just a stress test by the benchmark designers.
Is there any way to lift this limitation?
Best Regards, Peter
Hi Peter,
thanks for your hints. I've found a simple solution to increase the maximum number of attributes per element to 2^31, which should hopefully be enough (well, who knows?..). Please have a look at the latest stable version 6.3.4 (http://basex.org/download).
Feedback on your experiences with BaseX and the TPoX Benchmark is welcome! Christian _____________________________
2010/12/13 Peter Fischer peter.fischer@inf.ethz.ch:
Dear all,
when testing BaseX with the TPoX benchmark workload (http://tpox.sourceforge.net/) I could not load the input data set since it contains elements with more than 32 attributes.
The large number of attributes stems from the FIXML data (http://www.fixprotocol.org/specifications/fix4.4fixml) contained in the workload, so it is not just a stress test by the benchmark designers.
Is there any way to lift this limitation?
Best Regards, Peter -- Dr. Peter Fischer Systems Group/D-INFK/ETH Zurich peter.fischer@inf.ethz.ch Tel: +41/44/632 36 16 http://www.systems.ethz.ch/people/petfisch Universitätstr 6, CAB E75
On 12/14/2010 02:57 AM, Christian Grün wrote:
Hi Peter,
Hi Christian, all,
thanks for your hints. I've found a simple solution to increase the maximum number of attributes per element to 2^31, which should hopefully be enough (well, who knows?..). Please have a look at the latest stable version 6.3.4 (http://basex.org/download).
thanks for the quick fix. I can confirm that now the TPoX workload can be executed by BaseX 6.3.4.
It seems that the only remaining problem was the change in the order of parameters to the CREATE DB command. With versions 6.0 and 6.1 the document came before the database, now it is the other way. I could not find any explicit information in the documentation, so this might be something to point out in the change logs.
Feedback on your experiences with BaseX and the TPoX Benchmark is welcome!
Regards, Peter
Christian _____________________________
2010/12/13 Peter Fischer peter.fischer@inf.ethz.ch:
Dear all,
when testing BaseX with the TPoX benchmark workload (http://tpox.sourceforge.net/) I could not load the input data set since it contains elements with more than 32 attributes.
The large number of attributes stems from the FIXML data (http://www.fixprotocol.org/specifications/fix4.4fixml) contained in the workload, so it is not just a stress test by the benchmark designers.
Is there any way to lift this limitation?
Best Regards, Peter -- Dr. Peter Fischer Systems Group/D-INFK/ETH Zurich peter.fischer@inf.ethz.ch Tel: +41/44/632 36 16 http://www.systems.ethz.ch/people/petfisch Universitätstr 6, CAB E75
It seems that the only remaining problem was the change in the order of parameters to the CREATE DB command. With versions 6.0 and 6.1 the document came before the database, now it is the other way. I could not find any explicit information in the documentation, so this might be something to point out in the change logs.
...true; this change has only been documented in our subversion comments. The new syntax will remain stable for future versions, as well as other changes on the commands that have been performed after Version 6.0.
Christian
basex-talk@mailman.uni-konstanz.de