Search code examples
rdataframefactors

In R how to divide a row when a condition is met

I have a bunch of means associated with unique values. The unique values are lines that I tested on two different substrates. in R I want to build a loop that will take the unique value and divide the means (pheno_output) of substrate A by substrate B. and spit this into a new vector.

This is what I am working on. 

tmpmeans<-  rep(NA, 2);

for (i in unique(SplitnonMEANs$Line)){
  tmpmeans<- SplitnonMEANs$pheno_output[i]/SplitnonMEANs$pheno_output[i]
}

example data:

s;

Line phenotype FoodSource pheno_output
1    non        A           123
2    non        A           456
1    non        B           789
2    non        B           159

Can anyone help me do this?

Solution

  • If you only have two values per group, you can use any of these

    divide <- function(x) {
    stopifnot(length(x) == 2L)
    x[1L]/x[2L]
} 
aggregate(pheno_output ~ Line, df, divide)
#   Line pheno_output
# 1    1    0.1558935
# 2    2    2.8679245
with(df, tapply(pheno_output, Line, divide))
#         1         2 
# 0.1558935 2.8679245 
sapply(split(df$pheno_output, df$Line), divide)
#         1         2 
# 0.1558935 2.8679245