We’re delighted with the response to our webinar, Big Analytics for R Users without Big Hassles. Attendees were engaged and we didn’t have time to answer all the questions submitted through the chat window. We’ll answer the overflow questions in batches. Here’s the first batch.
- Is SciDB performance optimized for 2D arrays and 3D arrays?
Performance is optimized for multidimensional arrays, period. 2,3, 4, whatever.
- Can I write new algorithms for SciDB in R? Right now, we have to do everything in C++, but it’d be nicer to do in R.
We wanted to make SciDB easy to use from R so we started with the R front end. Support for running R functions from within database queries will be coming. We will support R-coded user-defined functions and aggregates for starters; but we are unlikely to support R-coded user-defined operators.
- Do you have any plans to create similar features (similar to what’s offered in SciDB-R), for Python numpy?
We don’t have a timetable, but we expect to do this. By the way, Bryan will be at the SciPy 2013 in Austin this year, so if you go, look him up and chat with him about features you’d like to see.
- Can SciDB be compiled on a Mac?
Linux is our supported platform but some users have compiled the SW for Macs and are running on Macs. Look for their comments on the SciDB forum.
- Do I need to know the scidb query language to use the SciDB and the R interface?
In some cases, yes. We have an initial set of tightly integrated operations, and for the rest, the query language is needed.
We will be continuously adding native R support for more operators. In practice a mix will be used for most serious problems.
Remember, the SciDB-R package is new. We will continue to add functionality regularly. Significant releases appear periodically on CRAN but the current development version is always available on github.
That’s all for now. Stay tuned for a subsequent post with more answers to overflow questions. And if you missed the webinar, watch this space. When the video recording of the webinar is available, we’ll put it on our website and announce it with a blog post.