orgtbl-aggregate: how to handle periods (.) in column header names

Question

I'm new to org mode and have been using Orgtbl Aggregate for tables. I import space-separated files as org tables, add an hline to make the first row into a header, name them with #+NAME (aside: is there a better way to name a table than to just write this tag by hand above it?) and then I use org-insert-dblock:aggregate to aggregate data in a new table below.

My header-names have periods in them because they are LaTeX pgfplots-friendly. I can't use spaces, dashes, or underscores. The MWE below shows that Aggregate doesn't like the periods, and exhibits unpredictable behaviour when I try escaped quotes.

Can/should aggregate dblock work with the periods (how) or should I simply clean my headers first (this is what I ended up doing)?

#+NAME: mwe-table                                                                 |
| A | b.b | cc  |                                                                 |
|---+-----+-----|                                                                 |
| 1 |   3 |   4 |                                                                 |
| 1 |   4 |   5 |                                                                 |
| 2 |   6 |   7 |                                                                 |
| 2 |   8 |   9 |                                                                 |
#+BEGIN: aggregate :table "mwe-table" :cols "A mean(b.b) mean('b.b') mean(\"b.b\")"
| A | mean(b.b)                 | mean('b.b')                    |   mean("b.b") ||
|---+---------------------------+--------------------------------+---------------||
| 1 | error(7, '"Expected `)'") | error(6, '"Expected a number") | 80.6666666667 ||
| 2 | error(7, '"Expected `)'") | error(6, '"Expected a number") | 80.6666666667 ||
#+END:                                                                            |
#+BEGIN: aggregate :table "mwe-table" :cols "A mean(cc) mean(bb) mean(\'b.b\')"   |
| A | mean(cc) | mean(bb)  | mean('b.b')                    |                     |
|---+----------+-----------+--------------------------------|                     |
| 1 |      4.5 | vmean(bb) | error(6, '"Expected a number") |                     |
| 2 |        8 | vmean(bb) | error(6, '"Expected a number") |                     |
#+END:                                                                            |

Answer 1

A new version of orgtbl-aggregate is available on Melpa and https://github.com/tbanel/orgaggregate. Column names can now contain arbitrary characters (space, dot, semicolon, etc.). Just quote them with single or double quotes.

#+NAME: funnynames
| first column | observed;number | computed/expected |
|--------------+-----------------+-------------------|
| a/experiment |             2.3 |               2.4 |
| a/experiment |            15.4 |              12.1 |
| b/test       |           -98.7 |               0.0 |
....

#+BEGIN: aggregate :table "funnynames" :cols "\"first column\" mean('observed;number');%.3f mean('computed/expected');%.4f" :cond (>= (string-to-number "observed;number") 0)
| 'first column' | mean('observed;number') | mean('computed/expected') |
|----------------+-------------------------+---------------------------|
| a/experiment   |                   8.633 |                    7.1333 |
| b/test         |                   3.350 |                    3.1500 |
#+END:

To install from Melpa, add this to your .emacs file:

(require 'package)
(add-to-list 'package-archives '("melpa" . "http://melpa.milkbox.net/packages/") t)

then type M-x package-list and search for orgtbl-aggregate.

Answer 2

Funny column names have just not been anticipated. This can be a useful new feature. Your proposed syntax (single or double quotes) seems the way to go. I will look at how to implement that.

(The unexpected behavior you see when computing mean(\"b.b\") is actually mean([98,46,98]), which happen to be the ascii codes of b . b. Of course, this is hardly the expected result).

The author of orgtbl-aggregate