Gabor Szarnyas

I'm actually interested in using Markdown for educational materials. I experimented with a lot of Markdown flavors, e.g.: - https://github.com/FTSRG/lecture-notes/wiki (GitHub Flavored Markdown for course syllabi & laboratories) - https://github.com/FTSRG/remo...

As we discussed on Slack, it should result in tuples like this, but this is a possible optimisation inside the engine - which could be supported by the compiler.

Queries work now, see 17f7d48077da3960275d3dfc7cc6f088671df652. The transformation are not very complex, so we can start testing these in the near future: https://github.com/FTSRG/trainbenchmark/tree/master/trainbenchmark-tool-neo4j/src/main/resources/transformations

We currently perform validation in the `BiValidationTest` class. Implementing an LDBC driver certainly would have its benefits, however, it cannot be used for benchmarking as it is. The reason for...

It's worth checking the Postgres docs: > If a query contains aggregate function calls, but no GROUP BY clause, grouping still occurs: the result is a single group row (or...

Related literature (thanks to @bergmanngabor): https://dl.acm.org/citation.cfm?id=137852

This causes BI Q7 to break. A simplified version shows the issue: ``` MATCH (message2:Message) OPTIONAL MATCH (message2:Message)

The problem can be pinpointed to the following condition in `AggregationNode`s: ```scala if (oldValues != newValues) ``` Of course, this line is there for a reason...

Currently, this breaks BI Q21. 

Note: the updates for the LDBC BI benchmarks are not yet worked out -- this is planned for Q3/Q4 2018.