Greetings. I'm sitting in on a weekly, informal, "brown-bag" seminar on data technologies in statistics. There are more people attending the seminar than there are weeks in which to give talks, so I may get by with being my usual, passive-slug self.
But I thought it might be useful to have a contingency plan and decided that giving a brief talk about Babel might be useful/instructive. I thought (and think) that mushing together (with attribution) some of the content of the paper [1] by The Gang of Four and the content of Eric's talk [2] might be a good approach. (BTW, if this isn't legal, desirable, permissible, etc., this would be a good time to tell me.) I liked the Pascal's Triangle example (which morphed from elisp to Python, or vice versa, in the two references), but I was afraid that the elisp routine "pst-check", used as a check on the correctness of the previously-generated Pascal's triangle, might be too esoteric for this audience, not to mention me. (The recursive Fibonacci function is virtually identical in all languages, but the second part is more obscure.) I thought it should be possible to use R to do the same sanity check, as R would be much more-familiar to this audience (and its use would still demonstrate the meta-language feature of Babel). Unfortunately, I haven't been able to find a way to communicate the output of the Pascal's Triangle example to an R source-code block. The gist of the problem seems to be that regardless of how I try to grab the data (scan, readLines, etc.) Babel always ends up trying to read a data frame (table) and I get an error similar to: <<<<<< > Error in scan(file, what, nmax, sep, dec, quote, skip, nlines, na.strings, > : line 1 did not have 5 elements Enter a frame number, or 0 to exit 1: read.table("/tmp/babel-3780tje/R-import-3780Akj", header = FALSE, row.names = NULL, sep = " >>>>>> If I construct a table "by hand" with all of the cells occupied, everything goes OK. For instance: <<<<<< #+TBLNAME: some-junk | 1 | 0 | 0 | 0 | | 1 | 1 | 0 | 0 | | 1 | 2 | 1 | 0 | | 1 | 3 | 3 | 1 | #+NAME: read-some-junk(sj_input=some-junk) #+BEGIN_SRC R rowSums(sj_input) #+END_SRC #+RESULTS: read-some-junk | 1 | | 2 | | 4 | | 8 | >>>>>> But the following gives the kind of error I described above: <<<<<< #+name: pascals_triangle #+begin_src python :var n=5 :exports none :return pascals_triangle(5) def pascals_triangle(n): if n == 0: return [[1]] prev_triangle = pascals_triangle(n-1) prev_row = prev_triangle[n-1] this_row = map(sum, zip([0] + prev_row, prev_row + [0])) return prev_triangle + [this_row] pascals_triangle(n) #+end_src #+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: pst-checkR(pas_inputs=pascals_triangle) #+BEGIN_SRC R rowSums(pas_inputs) #+END_SRC >>>>>> Note that I don't really want to do rowSums in this case. I'm just trying to demonstrate the error. Of course, it's clear that the first line does NOT contain five elements, nor does the second, etc., as all of the above-diagonal elements are blanks. But I've been unable to find an R input function that doesn't end up treating the source data as a table, i.e., in the context of Babel source blocks -- R is "happy" to read a lower-diagonal structure. See the appendix for an example. Any suggestions? Note that I'm happy to acknowledge that my own ignorance of R and/or Babel might be the source of the problem. If so, please enlighten me. Thanks. -- Mike [1] http://www.jstatsoft.org/v46/i03 [2] https://github.com/eschulte/babel-presentation <<<<<< appendix -------- $ cat pascal.dat 1 1 1 1 2 1 1 3 3 1 1 4 6 4 1 $ R --vanilla < pascal.R R version 2.15.0 (2012-03-30) Copyright (C) 2012 The R Foundation for Statistical Computing ISBN 3-900051-07-0 Platform: x86_64-redhat-linux-gnu (64-bit) . . . > x <- readLines("pascal.dat") > x [1] "1" "1 1" "1 2 1" "1 3 3 1" "1 4 6 4 1" > str(x) chr [1:5] "1" "1 1" "1 2 1" "1 3 3 1" "1 4 6 4 1" > > y <- scan("pascal.dat") Read 15 items > y [1] 1 1 1 1 2 1 1 3 3 1 1 4 6 4 1 > str(y) num [1:15] 1 1 1 1 2 1 1 3 3 1 ... > > z <- read.table("pascal.dat", header=FALSE) Error in scan(file, what, nmax, sep, dec, quote, skip, nlines, na.strings, : line 1 did not have 5 elements Calls: read.table -> scan Execution halted