[R] help on deleting NAs

[Patrick Connolly]

On Thu, 17-Feb-2005 at 02:54PM -0600, KeLin at mdanderson.org wrote:

|> Dear R friends
|> 
|> My goal is to eliminate this specific group(1) if the # of NAs in this 
|> group greater than
|> 50%(specifically say greater than 3). Would you please show me how to do 
|> it. 
|> I have a sample data as following:
|> 
|> Thanks a lot.
|> 
|> Kevin Lin
|> 
|>            y group f1 f2 f3
|> 30         NA     1  1  1  1
|> 27         NA     1  1  2  2
|> 48         NA     1  2  1  2
|> 40 -0.6066416     1  2  2  1
|> 24 -0.8323225     1  3  2  2
|> 25  1.3401226     2  1  1  1
|> 13  1.2619082     2  1  2  1
|> 14 -0.4323220     2  3  1  1
|> 36  0.8406529     2  3  2  2
|> 21  0.9604758     3  1  2  1
|> 18  0.9562072     3  2  1  1
|> 45  1.1285016     3  2  1  1
|> 50         NA     4  1  1  1
|> 11         NA     4  1  1  2
|> 41 -1.1017167     4  2  1  1
|> 37  0.9661283     4  3  1  1
|> 39 -0.2540905     4  3  1  2


There's probably a lot of niftier ways but this will give an idea:
If X is your dataframe above,

> aa <- with(X, tapply((y), group, function(x) length(x[is.na(x)])))
> names(aa[aa>2])
[1] "1"

> X[!with(X, group%in%as.numeric(names(aa[aa>2]))),]
            y group f1 f2 f3
6   1.3401226     2  1  1  1
7   1.2619082     2  1  2  1
8  -0.4323220     2  3  1  1
9   0.8406529     2  3  2  2
10  0.9604758     3  1  2  1
11  0.9562072     3  2  1  1
12  1.1285016     3  2  1  1
13         NA     4  1  1  1
14         NA     4  1  1  2
15 -1.1017167     4  2  1  1
16  0.9661283     4  3  1  1
17 -0.2540905     4  3  1  2
> 

The function in the tapply part could be made more general if 3
doesn't always constitute a majority.

HTH

-- 
Patrick Connolly
HortResearch
Mt Albert
Auckland
New Zealand 
Ph: +64-9 815 4200 x 7188
~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~
I have the world`s largest collection of seashells. I keep it on all
the beaches of the world ... Perhaps you`ve seen it.  ---Steven Wright 
~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~