> If I add fill=TRUE to that (on a git branch), then I get this: > > #+RESULTS: pascals-triangle > | 1 | | | | | | > | 1 | 1 | | | | | > | 1 | 2 | 1 | | | | > | 1 | 3 | 3 | 1 | | | > | 1 | 4 | 6 | 4 | 1 | | > | 1 | 5 | 10 | 10 | 5 | 1 | > > #+NAME: sanity-check > #+HEADER: :var sc_input=pascals-triangle > #+BEGIN_SRC R > sc_input > #+END_SRC > #+RESULTS: sanity-check > > | 1 | nil | nil | nil | nil | > | 1 | 1 | nil | nil | nil | > | 1 | 2 | 1 | nil | nil | > | 1 | 3 | 3 | 1 | nil | > | 1 | 4 | 6 | 4 | 1 | > | 1 | 5 | 10 | 10 | 5 | > | 1 | nil | nil | nil | nil | > > which isn't correct, but gets past the scan error. >
Hmm, this happens with my patch applied as well. It seems to me this *must* be an R error. The raw textual data pre-import has no such wrap.
1 1 1 1 2 1 1 3 3 1 1 4 6 4 1 1 5 10 10 5 1
Why would R intentionally wrap a table at an arbitrary column? > > I'm in over my head here, but hope that my curiosity hasn't been too > noisy. > Me too. Unless someone who is familiar with the motivations and design decisions behind R's read.table function, I'm inclined to leave the current Org-mode code as is. Thanks, -- Eric Schulte http://cs.unm.edu/~eschulte/