R - Closure subsetting error during loop

Question

I have the following code to calculate ADF statistics across ts data in a matrix raw.

raw <- matrix(rnorm(25000), nrow=1000,ncol=25)
for (i in 1:(ncol(raw)-1)) {
    for (j in (i+1):ncol(raw)) {

        df <- data.frame(N = integer(0), DFStat = numeric(0), PVAL =     numeric(0))
        name <- paste(i,j, sep="_")
        dir <- paste("C:/Data/rOUT/", name,".txt",sep="")
        for (t in 1:20) {
             N <- 50*t
             xtail <- tail(raw[,i],N)
             ytail <- tail(raw[,j],N)
             reg <- lm(ytail ~ xtail)
             test <- adf.test(reg$residuals, k = 1)
             df[t,] <- c(N, test$statistic, test$p.value)
             write.table(df, dir, sep="\t", row.names=FALSE)
             rm(df)

        }
    }
}

The block in the first two loops runs perfectly when not looping through the columns in raw, but when I try to traverse raw, I get the closure error below:

Error in df[t, ] <- c(N, test$statistic, test$p.value) : 
  object of type 'closure' is not subsettable

Does anyone know why I am encountering this difference and steps I can take to correct it?

Your code doesn't do anything without data, so it's practically impossible to debug without it. You should provide minimal code to reproduce your problem (see http://stackoverflow.com/help/mcve and http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) — Phil, Jan 18 '17 at 13:20
Have a look at class(test$statistic) and class(test$p.value). Is it a number which fits in your df? — Christoph, Jan 18 '17 at 13:48
@christoph yes the tests return a Dickey-Fuller score and a p.value, both of which are numerics as defined when i initialized the data frame. I think R is treating df as a function when I run it as the loop, I have no idea why though. — Ray Foo, Jan 18 '17 at 16:26

score 1 · Accepted Answer · answered Jan 18 '17 at 16:52

I guess your rm(df) is at the wrong position. The following code runs:

raw <- matrix(rnorm(250), nrow=10,ncol=25)
for (i in 1:(ncol(raw)-1)) {
  for (j in (i+1):ncol(raw)) {

    df <- data.frame(N = integer(0), DFStat = numeric(0), PVAL =     numeric(0))
    name <- paste(i,j, sep="_")
    dir <- paste("C:/Data/rOUT/", name,".txt",sep="")
    for (t in 1:20) {
      N <- 50*t
      xtail <- tail(raw[,i],N)
      ytail <- tail(raw[,j],N)
      reg <- lm(ytail ~ xtail)
      # test <- adf.test(reg$residuals, k = 1)
      # df[t,] <- c(N, test$statistic, test$p.value)
      df[t,] <- c(N, 1, 1)
      # write.table(df, dir, sep="\t", row.names=FALSE)


    }
    rm(df)
  }
}

Ah you're right, late night coding errors I suppose. Thanks! — Ray Foo, Jan 18 '17 at 16:56

R - Closure subsetting error during loop

1 Answers1