Lately I think Mac on Travis should all be an allowed failure because those queues are pretty much broken on Travis’s end …
2 Likes
Great, good to hear that ODBC works for you.
For the record, JDBC has a readtable method that takes a ResultSet and returns a DataFrame. I realise now that this function is not clearly documented. Need to fix that, thanks!
Regards
Avik
That’s why my tests usually have println("<HEARTBEAT>") all over the place 
1 Like
I’m a major user of DataFrames and a minor contributor to it. I can tell you that the work being done is massive, and of excellent quality. Its going to be a top package when Julia 1.0 hits, and is already extremely useful. I used it on big projects involving billions of records.
Escher did not work for me when I wanted to use it, but it’s in development.
Debugging did not work until I went to JuliaPro. It may be broken again. Its up against LLVM changes at the moment, I think. When Julia 1.0 hits I am reasonably certain it will be stable again. Lots of work is being done on it.
Deployment currently requires a bit more work than some .Net application you could build, as an example. It can be pre-compiled. It will get better with or just after Julia 1.0.
I stopped using R and Python in 2016 and switched over to Julia. The painful bits have been worth the improvement in the coding experience and the application performance.
4 Likes
Thanks a lot for your experience report.
This is definitely what I was hoping for.
Could you explain what you mean by “improved coding experience”?
R’s syntax and language feel quite dated to me. Julia incorporates so many modern concepts, programming in that language is a vastly improved experience. YMMV
1 Like
PostgreSQL.jl is dead, but I have the go-ahead to work on LibPQ.jl. Expect DataStreams.jl integration in less than a month (I’m on vacation for a while).
We’re currently using ODBC.jl but it’s proven challenging to set up the backing library and get connections working in a variety of environments.
5 Likes