Question

I have a data set raw.data.2010 that needs several steps of subsetting with different animal species. I also need to name them accordingly after every filtering process. I wrote a simple code as below:

#Creating reproducible data######
site=rep(list("Q", "R", "S", "T"), each=500)
grid=sample(1:2, size=2000, replace=TRUE)
spp=rep(list("A", "B", "C", "D", "E"), each=400)
fate=sample(1:5, size=20000, replace=TRUE)
sex=rep(list("M","F"), each=2000)
weight=sample(85:140, size=2000, replace=TRUE)

raw.data=as.data.frame(cbind(site, grid, spp, fate, sex, weight))

### main codes#####
spp=c("A", "B", "C", "D", "E")
    for (i in spp){
        name=paste(i, "raw", sep=".", collapse="")
        filter=paste("select",i, sep="", collapse="")
        assign(filter, raw.data$spp==i)
        assign(name, raw.data[get(filter),])
    }

I checked the filters and they worked without problem. But the last line didn’t work so all the subsetted data I called returned NA. What was wrong? Thank you.

EDIT: Hi, thank you all for your advice. I edited my codes so it’s reproducible. Basically I would like to first filter my raw.data with spp. Then I can keep adding more filters to group them according to site, grid, fate…etc. I need to be able to access the filtered data individually so I can manipulate them for later use, ex. calculate weight and other measurements for different sex or age group. I want to be able to call A.raw, A.Q.data later.

Since I would like to analyze my data at different levels (e.g. population level, individual level, site/grid level), and be able to pool/split them according to my needs. That’s the purpose of this code. Hope my explanation doesn’t confuse you.

Answer 1

You will probably save yourself a lot of work and grief in the long run if you move away from using global variables with assign and get and instead work with lists (and remember to subset using [[ instead of $).