Subtraction within Groups using R

Question

I have been struggling to find the best way to do this subtraction within groups. I have data frame containing a long list of samples (Sample), each one treate

Answer 1

Try this

> transform(df, dValue=ave(value, Sample, FUN=function(x) c(0, x[1]-x[-1])))
  Sample condition value dValue
1   var1         A    12      0
2   var1         B    14     -2
3   var1         C    15     -3
4   var2         A    20      0
5   var2         B    19      1
6   var2         C    19      1
7   var3         A    50      0
8   var3         B    51     -1
9   var3         C    48      2

I'm assuming df looks like this

  Sample condition value
1   var1         A    12
2   var1         B    14
3   var1         C    15
4   var2         A    20
5   var2         B    19
6   var2         C    19
7   var3         A    50
8   var3         B    51
9   var3         C    48

Answer 2

You can also do that with dplyr:

require(dplyr)

df %.% 
   group_by(Sample) %.% 
   mutate(dValue = value[condition == "A"] - value)

#  Sample condition value dValue
#1   var1         A    12      0
#2   var1         B    14     -2
#3   var1         C    15     -3
#4   var2         A    20      0
#5   var2         B    19      1
#6   var2         C    19      1
#7   var3         A    50      0
#8   var3         B    51     -1
#9   var3         C    48      2

Answer 3

Assuming your dataset is called dat, here's a couple of data.table solutions:

Method 1:

require(data.table) ## >= 1.9.2
setDT(dat)[, dValue := value[condition == "A"] - value, by=Sample]

Method 2:

require(data.table) ## >= 1.9.2
setkey(setDT(dat), Sample)
dat[dat[condition == "A"], dValue := i.value-value]

#    Sample condition value dValue
# 1:   var1         A    12      0
# 2:   var1         B    14     -2
# 3:   var1         C    15     -3
# 4:   var2         A    20      0
# 5:   var2         B    19      1
# 6:   var2         C    19      1
# 7:   var3         A    50      0
# 8:   var3         B    51     -1
# 9:   var3         C    48      2