R help - Jul 2011 - Getting wrong NA values using "for" cmd

Hi There,

I'm facing one problem to construct a vector using the "for"
command:

I have one matrix named 'dados' (same as /data/ from portuguese), for
example:
> dados[140:150,]
          [,1] [,2] [,3]
 [1,] 212.7298 0.14 0.11
 [2,] 213.3778 0.14 0.11
 [3,] 214.0257 0.15 0.11
 [4,] 214.6737 0.15 0.12
 [5,] 215.3217 0.15 0.12
 [6,] 215.9696 0.15 0.12
 [7,] 216.6176 0.16 0.12
 [8,] 217.2656 0.16 0.13
 [9,] 217.9135 0.16 0.13
[10,] 218.5615 0.16 0.13
[11,] 219.2094 0.17 0.13

So, I need one vector getting the values from the third column given
specific values of the second:

s <- seq(0,1,0.05)
r <- NULL
for (w in 1:length(s)){
r[w] <- dados[,3][dados[,2]==s[w]][1]
}

This vector 'r' are working well to many values, but there are some
inconsistency (NA's) for others:
> r
 [1] 0.00 0.03 0.07   NA 0.16 0.21   NA   NA 0.36 0.41 0.46 0.52   NA 0.63  
NA
[16] 0.74 0.79   NA 0.90   NA 1.00

and dissecting this:
> s[4]
[1] 0.15
> dados[,3][dados[,2]==0.15][1]
[1] 0.11
> dados[,3][dados[,2]==s[4]][1]
[1] NA
> s[4]==0.15
[1] FALSE
> s[5]==0.20
[1] TRUE
> dados[,3][dados[,2]==s[5]][1]
[1] 0.16

Anyone could help me to discover why this is occorring?

dados[,2] and dados[,3] came from rounded values. Could this fact interfere
the results? 

Thanks in advance.

--
View this message in context:
http://r.789695.n4.nabble.com/Getting-wrong-NA-values-using-for-cmd-tp3654335p3654335.html
Sent from the R help mailing list archive at Nabble.com.

Looks like FAQ 7.31 to me.

Try all.equal() instead of ==.

Sarah

On Fri, Jul 8, 2011 at 11:06 AM, VictorDelgado
<victor.maia at fjp.mg.gov.br> wrote:
> Hi There,
>
> I'm facing one problem to construct a vector using the "for"
command:
>
> I have one matrix named 'dados' (same as /data/ from portuguese),
for
> example:
>
>> dados[140:150,]
> ? ? ? ? ?[,1] [,2] [,3]
> ?[1,] 212.7298 0.14 0.11
> ?[2,] 213.3778 0.14 0.11
> ?[3,] 214.0257 0.15 0.11
> ?[4,] 214.6737 0.15 0.12
> ?[5,] 215.3217 0.15 0.12
> ?[6,] 215.9696 0.15 0.12
> ?[7,] 216.6176 0.16 0.12
> ?[8,] 217.2656 0.16 0.13
> ?[9,] 217.9135 0.16 0.13
> [10,] 218.5615 0.16 0.13
> [11,] 219.2094 0.17 0.13
>
> So, I need one vector getting the values from the third column given
> specific values of the second:
>
> s <- seq(0,1,0.05)
> r <- NULL
> for (w in 1:length(s)){
> r[w] <- dados[,3][dados[,2]==s[w]][1]
> }
>
> This vector 'r' are working well to many values, but there are some
> inconsistency (NA's) for others:
>
>> r
> ?[1] 0.00 0.03 0.07 ? NA 0.16 0.21 ? NA ? NA 0.36 0.41 0.46 0.52 ? NA 0.63
> NA
> [16] 0.74 0.79 ? NA 0.90 ? NA 1.00
>
> and dissecting this:
>
>> s[4]
> [1] 0.15
>> dados[,3][dados[,2]==0.15][1]
> [1] 0.11
>> dados[,3][dados[,2]==s[4]][1]
> [1] NA
>> s[4]==0.15
> [1] FALSE
>> s[5]==0.20
> [1] TRUE
>> dados[,3][dados[,2]==s[5]][1]
> [1] 0.16
>
> Anyone could help me to discover why this is occorring?
>
> dados[,2] and dados[,3] came from rounded values. Could this fact interfere
> the results?
>
> Thanks in advance.
>
> --
-- 
Sarah Goslee
http://www.functionaldiversity.org