[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Potential student projects

For Avatica, there are a few issues open for better integration with HSQLDB (support for more types and some bugs):
- CALCITE-1951: https://issues.apache.org/jira/browse/CALCITE-1951
- CALCITE-2250: https://issues.apache.org/jira/browse/CALCITE-1950
- CALCITE-1957: https://issues.apache.org/jira/browse/CALCITE-1957

Another interesting one is to implement client-driven load balancing into the Avatica clients (see https://community.hortonworks.com/articles/9377/deploying-the-phoenix-query-server-in-production-e.html). I think this one would be quite interesting and would be something that requires a bit of design and thinking.

The avatica code base is not too big and shouldn't be too hard to understand. The downside is that the work is not as interesting as the ones available in Calcite, but I'll see if I can find some interesting other interesting issues for Avatica too.


On 14/11/2018 6:46 am, Michael Mior wrote:
Agreed. Students only have a few months to take this project from start to
finish so anything which helps them get up to speed is a plus.

Michael Mior

Le mar. 13 nov. 2018 à 14:39, Julian Hyde <jhyde@xxxxxxxxxx> a écrit :

I think it’s important that the project should start off with some
“mechanical” stuff that doesn’t require any design. Coming up to speed on a
new code base is challenging. The GIS functions work has that property.

On Nov 13, 2018, at 11:36 AM, Michael Mior <mmior@xxxxxxxxxx> wrote:

Good thought. I'll keep that in mind as well. There's probably far more
projects (and students) than I'll be able to supervise, but hopefully
Calcite will be able to get some mileage from this.

Michael Mior

Le mar. 13 nov. 2018 à 13:04, Julian Hyde <jhyde@xxxxxxxxxx> a écrit :

There’s always more work to be done in Geospatial, e.g.
https://issues.apache.org/jira/browse/CALCITE-2031 <
https://issues.apache.org/jira/browse/CALCITE-2031>. The easy part is
create scalar (row-at-a-time) implementations of spatial functions, but
real fun starts when you transform queries to use materialized views
(“indexes”). Until now, everything has been on rectangular grids and
space-filling curves (i.e. assuming flat-earth) but indexing using Uber
[1] or Google S2 would be challenging and cutting-edge.


[1] https://github.com/uber/h3

On Nov 12, 2018, at 9:02 AM, Michael Mior <mmior@xxxxxxxxxx> wrote:

Hi all,

One of the perks of being a career academic is the availability of
who are looking for interesting projects to work on. I'll be having
students work with me on a regular basis on final projects for their
degree. I'm hoping that some of these can be useful to Calcite.

If anyone can think of a few issues that are technically interesting,
too large in scope, and don't currently have anyone working on them,
all ears :)

A few issues below I found which could be interesting although some
be too large in scope for someone who's never seen Calcite before.


Michael Mior