Allow DataSet to operate on already-parsed table data

[GoogleCodeExporter]

I am looking at using Axiis as the graphing library for an (open-source)
Air application for to visualize Grinder data files.

While Grinder does use CSV format for its data files, the contents of the
file really need to be processed before that information is ready to be
graphed.  For example, rather than plot every Grinder agent sample point
(potentially hundreds of thousands), I will prepare a number of aggregating
sample points and then graph those instead.  I would like to avoid parsing
the CSV and then *re-creating* a simplified CSV just so that it can be
parsed by Axiis.

I propose some method that can be given (already-parsed) headers and rows
and will create its internal rows/columns/etc. structure based on these
structures.  (Additionally it would add any aggregate stuff at that point.)
 Maybe processTable() -- though maybe a factory pattern makes more sense
here ... ?   DataSet.tableFromArrays(headers:Array, rows:Array) or
DataSet.tableFromCSV() ?

My thought would be that the processTableFromCsv() could use this method
under the hood too.  Anyway, I'm happy to propose a patch for this, though
it probably merits some further on-list discussion first.

Original issue reported on code.google.com by h...@velum.net on 2 Aug 2009 at 3:48