CODEWITHSUNDEEP
rdataframesubset
neversaint
neversaint

Reputation: 64024

How to deal with missing row lines when performing subset in R

I have a data that looks like this

   Cov dis err kval  n         prob
1   10   0   0    0  0 7.574657e-01
2   10   0   0    0  1 2.104075e-01
3   10   0   0    0  2 2.922241e-02
4   10   0   0    1  3 2.705617e-03
5   10   0   0    1  4 1.878731e-04
6   10   0   0    0  5 1.043613e-05
7   10   0   0    0  6 4.830806e-07
8   10   0   0    0  7 1.916636e-08
9   10   1   0    0  8 6.653565e-10
10  10   1   0    0  9 2.053068e-11
11  10   1   0    0 10 5.701398e-13

What I want to do is perform a subset on that data. But why this failed?

dat.full <- read.table("http://dpaste.com/1283733/plain/",header=TRUE);
subset(dat.full,(Cov == 10    && dis==0 &&  err==0 && kval==1), select=prob);

It gave the following missing row output:

[1] prob
<0 rows> (or 0-length row.names)

What's the correct way to perform the subset?

I expect it to return:

4 2.705617e-03
5 1.878731e-04

Upvotes: 0

Views: 62

Answers (2)

Beasterfield
Beasterfield

Reputation: 7113

You can fix this with subsetting by

subset(dat.full, (Cov == 10 & dis==0 & err==0 & kval==1), select=prob)

The problem you bumped in is, that && is not vectorized but should be only used for scalar values. Use & (or | for OR) instead.

Upvotes: 1

agstudy
agstudy

Reputation: 121578

You should use & and not && for vectorized operations.

 subset(dat.full ,Cov == 10    & dis==0 &  err==0 & kval==1,select=prob)
          prob
4 0.0027056170
5 0.0001878731

Upvotes: 3

Related Questions