R help - Apr 2011 - subset without removing NAs

Hi folks,

I'm dealing with a dataset with a lot of NAs, and want to use subset on the
data without removing the NAs from the the dataframe; e.g.

rws50 <- subset(rw.fire.RW,shigo.av<50)

This removes instances with NA for the value in addition to anything <50.
Any suggestions much appreciated.

*Ben Caldwell*

	[[alternative HTML version deleted]]

On 04/28/2011 09:53 PM, Benjamin Caldwell wrote:
> rws50<- subset(rw.fire.RW,shigo.av<50)
quick and dirty would be to replace all NAs with -99999 (or similar), 
use subset, and set all values ==-99999 in the subset back to NA. There 
may be more elegant solutions, though.


Jannis

On Apr 28, 2011, at 3:21 PM, Jannis wrote:
> On 04/28/2011 09:53 PM, Benjamin Caldwell wrote:
>> rws50<- subset(rw.fire.RW,shigo.av<50)
> 
> quick and dirty would be to replace all NAs with -99999 (or similar), use
subset, and set all values ==-99999 in the subset back to NA. There may be more
elegant solutions, though.
> 

Indeed. See ?is.na

DF <- data.frame(A = 1:20, B = c(1:15, rep(NA, 5)))
> subset(DF, B > 10)
    A  B
11 11 11
12 12 12
13 13 13
14 14 14
15 15 15
> subset(DF, (B > 10) | is.na(B))
    A  B
11 11 11
12 12 12
13 13 13
14 14 14
15 15 15
16 16 NA
17 17 NA
18 18 NA
19 19 NA
20 20 NA


HTH,

Marc Schwartz

On Apr 28, 2011, at 3:53 PM, Benjamin Caldwell wrote:
> Hi folks,
>
> I'm dealing with a dataset with a lot of NAs, and want to use subset  
> on the
> data without removing the NAs from the the dataframe; e.g.
>
> rws50 <- subset(rw.fire.RW,shigo.av<50)
>
> This removes instances with NA for the value in addition to anything  
> <50.
> Any suggestions much appreciated.
If you use "[" instead the NA's will come though:

rw.fire.RW[ shigo.av<50, ]

In fact I generally wrap my logicals in which to avoid such  
(mis?)behavior. but you appear to approve of it.

=David Winsemius, MD
West Hartford, CT