I'd like to implement the Clickhouse[1] bridge, perhaps I can find some time in the near future. There is a client library[2] which quiet nicely aligns with Arrow's columnar format. I'd also consider MySQL, because that's the most popular database.
[1] clickhouse.yandex [2] https://github.com/artpaul/clickhouse-cpp On Tue, Feb 19, 2019 at 5:16 PM Wes McKinney <wesmck...@gmail.com> wrote: > I agree with Antoine. The more well-defined and less uncertain the > project, the higher the probability of success. I had suggested > implementing a bridge between one or more database protocols (e.g. > SQLite3 or libpq / PostgreSQL) as example projects that could get done > in 3 months. By the way, if anyone is interested in working on these > projects independent of GSoC please reach out to me. > > - Wes > > On Tue, Feb 19, 2019 at 3:24 AM Antoine Pitrou <anto...@python.org> wrote: > > > > > > Le 19/02/2019 à 03:59, Tanya Schlusser a écrit : > > > Would developing an open standard for in-memory records qualify as > 'GSoC' > > > worthy? > > > > > > In reference to this placeholder in the Confluence wiki: > > > > > > > https://cwiki.apache.org/confluence/display/ARROW/Apache+Arrow+Home#ApacheArrowHome-Developinganopenstandardforin-memoryrecords > > > which links to ARROW-1790 > > > https://issues.apache.org/jira/browse/ARROW-1790 > > > and to this thread > > > > > > > https://lists.apache.org/thread.html/4818cb3d2ffb4677b24a4279c329fc518a1ac1c9d3017399a4269199@%3Cdev.arrow.apache.org%3E > > > > > > Developing a standard, or even just starting a standard working group > would > > > be quite a contribution, and allow a grad student the opportunity to > > > contact multiple leaders in the field. (I am thinking of something > along > > > the lines of the Data Mining Group http://dmg.org/, which I believe > is run > > > by a local professor here in Chicago). > > > > My indirect experience (I have not mentored a GSoC student, but I have > > followed projects who had GSoC students at some point) is that GSoC > > projects must be focussed enough, and there should be little to no > > unknowns, so that the student can progress without getting lost. So I > > don't think asking to develop or start designing a standard is a good > idea. > > > > Of course there may be the occasional brillant student who's able to > > overcome all that. > > > > Regards > > > > Antoine. >
