By combining transpose and rank you can apply a function to almost any
subset of
of the data in an array. Transpose provides the abililty to reorder the
axes in any
way you like. Is that what you are wanting to do?
Hadley Wickham's paper you quote appears to want to create new objects based
on using values in the data set. J's key adverb facilitates that.
For an isigraph implementation of pivot tables look at multi-dimensional
arrays in the 'grid' demo.
Fraser
----- Original Message -----
From: "June Kim" <[EMAIL PROTECTED]>
To: "Programming forum" <[email protected]>
Sent: Monday, June 09, 2008 2:44 PM
Subject: [Jprogramming] R's reshape
I've been thinking pivot tables are a major innovation in the history
of spreadsheet. I wondered if there is a natural way of pivot-tabling
in more textual/literal environments like J.
I think I've found something. It's R's reshape. http://had.co.nz/reshape/
You may read the pdf paper that's linked on the page
http://www.jstatsoft.org/v21/i12
R is a array-oriented programming language, hence a descendant of APL,
which is mostly used in statistics domain.
What do you think about reshape? How would J do pivot-tabling? Would
it be similar to reshape or different?
June
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
