[GRASS-dev] Re: [GRASS GIS] #430: i.pca's output and history logged "Comments" disagree

Markus Neteler neteler at osgeo.org
Wed Mar 18 18:08:31 EDT 2009

On Tue, Mar 17, 2009 at 12:31 PM, Hamish <hamish_b at yahoo.com> wrote:
> Markus:
>> Patch attached (maybe there is a better way to write it):
>> Eigen values, (vectors), and [percent importance]:
>> PC1 4712.17 (-0.2798,-0.3332,-0.5054,-0.0015,-0.5341,-0.5196) [ 84.0537% ]
> ....
> just cosmetics, but please could it look like:
> (G_message version)
>> PC1 4712.17 ( -0.2798, -0.3332, -0.5054, -0.0015, -0.5341, -0.5196 ) [ 84.05% ]
> ie keep percentage as %.2;


> keep a space between vector components;

<strike>why that? for readability ? (it's not CSV format conform) </strike>

after reading it all I perhaps got the point.

> and (less importantly)  keep the space before and after the whole vector component set.

shrug ok.

> the idea is that the increased component prec is needed so you can reconstruct the PCs,
> but percentage is only informational so it needs less (& you can recalc it from the eigenvalues
> if you want more). The r.info -h version should have it all align in a nice table so that you
> can column-wise import it as fixed-width into a speadsheet

ah, will check that

> or box-select the values in a nice text editor that allows that (nedit, wordstar, etc). without the
> extra whitespace it is a bit too cluttered for my liking.


> G_message() compacts whitespace so you don't see the nice formatting in the normal
>  output, and is a byproduct so shouldn't be sent to stdout by default.

ok... as time permits. If anyone else is faster: welcome!


More information about the grass-dev mailing list