Bug introduced in 10.3

I am using JoinAcross to combine two datasets, where the key is an Integer.

L = Get["https://dl.dropboxusercontent.com/u/557392/permalink/StackExchange/mathematica/107732/L.m"];
R = Get["https://dl.dropboxusercontent.com/u/557392/permalink/StackExchange/mathematica/107732/R.m"];

JoinAcross[L, R, "A", "Inner"]

The Dataset rendition of these associations is

For L:

L Dataset

For R:

R Dataset

Though this takes a couple of seconds in KNIME and/or GNU R, it takes an hour or more in Mathematica 10.3.1 on GNU/Linux.

What can I do to speed this operation up?

  • 2
    $\begingroup$ Ask Konstantin at WRI, he wrote that I believe. $\endgroup$ – M.R. Feb 19 '16 at 2:09
  • 3
    $\begingroup$ 4.2 seconds, v10.1/Windows . Similar on my linux box. $\endgroup$ – george2079 Feb 19 '16 at 22:10
  • 2
    $\begingroup$ 3575.66 seconds, v10.3.1/Windows $\endgroup$ – Alexey Golyshev Feb 20 '16 at 14:11
  • 2
    $\begingroup$ 3.0 seconds v10.2/Mac. I guess we've bracketed it to a change between 10.2 and 10.3. Thanks everyone, I have contacted Wolfram Support. $\endgroup$ – Eric Brown Feb 22 '16 at 14:14
  • $\begingroup$ This is a thousand-fold slowdown. I think the bugs tag was warranted. $\endgroup$ – Szabolcs Mar 2 '16 at 18:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.