[R] Subsetting with more than one criteria

[Prof Brian Ripley]

On Fri, 29 Oct 2004, CG Pettersson wrote:

> I have been using R for two years, but still have trivial problems.
> For the moment, I run R1.9.1 on a W2k computer.

Time to read a book!  This sort of thing is discussed at length in Chapter
2 of MASS4 (see the FAQ), for example.  It even discusses why your example
does not work.

> I work with a dataset from a variety evaluation project. Every year 25
> varieties have been tested. The tested cultivars have been the same
> each year in all trials, but some have changed over the three year
> period. Now I want to do some calculations and plotting on subsets.
> The structure looks like this:
> 
> Year	ADB	Block	Vcode	Variety	Yield	Protein			
> 2002	SW0024	1	20226	Denise	5843	12.8....	
> 2002	SW0024	1	9865	Astoria	6729	11.4	
> 2002	SW0024	1	9622	Barke	6121	12	
> 2002	SW0024	1	9604	Cecilia	5579	12.7	
> 2002	SW0024	1	20223	Granta	5591	11.6	
> 2002	SW0024	1	20222	Class	5591	11.7	
> 2002	SW0024	1	9922	Wiking	5744	12.5
> 2002	SW0024	1	20103	Vortex	5863	10.6
> .
> .
> And so on both down and sideways.
> Three years and four trials with three replicats each gives 900 lines.
> 
> How do I use several criteria to subset this?
> 
> Ast <- subset(data, Variety == "Astoria") 
> 
> works fine giving back the 36 lines where Astoria appears.
> But how do I pick two (or more) varieties? AND picking one (or two)
> years?
> 
> Everything I have tried results in rubbish, like:
> 
> AstBark <- subset(data, Variety == c("Astoria","Barke"))

Variety %in% c("Astoria","Barke") is probably what you intended.

You can pick also on years by using & in the indexing criterion.
 
-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595