Skip to end of metadata
Go to start of metadata

Concall to discuss about the generation of test data

Participants: Norbert, Alex, Mirko, Renzo
lead: Norbert

Renzo:
propose a file with the following structure in XML

<samples>
 <persons maxID="5000"/>
 <posts from="2005-01-01T12:00:00Z" to="2006-12-31T12:00:50Z"/>
 <workFrom from="2004-12-31T12:00:50Z" to="2007-01-01T12:00:00Z">
</samples>

Mirko:
comments some changes in the specification of the queries and the requirements for test data.

Alex: it is necessary that the DBGEN provides informations about the distributions.
For example, how frequent a location is.

Norbert. It will be considered as a future improvement.

Norbert:
recommends JSON instead of XML
one single file with all the information
including the data from the dictionaries (e.g., countries)

All:
we agree the use of JSON

Mirko:
check the following files to select to test data
http://www.ldbc.eu:8090/display/TUC/Substitution+parametars
https://github.com/ldbc/ldbc_socialnet_bm/tree/master/ldbc_socialnet_qgen/td_data

Alex:
comments about the CSV files.


Norbert:
We return with this on september.
We need the queries at the end of august.
Virtuoso and Dex must run the benchmark on September.

All
holidays: Norbert 3 weeks, Alex available, Mirko two weeks, Renzo available.

  • No labels