Re: gnumeric is slow on certain operations

[jcupitt gmail com]

On 13 July 2011 11:28, Frederic Parrenin <parrenin gmail com> wrote:
> I put a sample spreadsheet here:
> http://parrenin.frederic.free.fr/download/bugzilla/Ice-synchros-2011-07-07.gnumeric

This is a very large spreadsheet. You have sheets of data with 210264,
266106 and 163349 rows, then the same data is processed several times
in other sheets as well. My gnumeric (also 1.10.13) opened it, but had
a great many NAME errors. Excel and oocalc were unable to open the
spreadsheet at all.

I wonder if a spreadsheet is the right tool for data analysis on this
scale. I find it difficult to have confidence in a 1000 row
spreadsheet, I can't really imagine working with a larger set like
this.

In my experience the best solution is to develop ideas experimentally
in a spreadsheet using a small subset of data, then reformulate as a
tiny Ruby or Python program for large-scale analysis (no doubt
R/Octave/Matlab would work as well). I analyse hundreds of megabytes
of proteomic data pretty easily this way. Probably not a helpful
reply, sorry :-(

John