CODEWITHSUNDEEP
rplotpnglapplypar
emr2
emr2

Reputation: 1722

How to save several plots generated from par(mfrow) and lapply?

I am trying to save in my pc all the plots obtained using par(mfrow). The idea is to save each png image adding a number to avoid replacements.

This is a toy example were I use lapply and par(mfrow) to generate several histograms.

par(mfrow=c(1,2))
  lapply(1:ncol(mtcars), function(x){
    hist(mtcars[,x])
  })

Something that I have tried (without success) is to use another lapply to save the images with different names.

lapply(1:ceiling(ncol(mtcars)/2), function(n){
  
  png(paste0("../Desktop/histogram_", n, ".png"))
  
  par(mfrow=c(1,2))
  lapply(1:ncol(mtcars), function(x){
    hist(mtcars[,x])
  })
  dev.off()
})

It works at 50%. It saves the different histograms with different names, but it saves the last plot several times.

How can I save the first plot till the last one?

Important note: I don't want to save each individual histogram, I want to save each pair of histograms that are kept in each image.

I found a similar post to do this, but it uses a for loop and if statements (something that I want to avoid).

Any help will be very appreciated.

Thanks in advance

Upvotes: 0

Views: 279

Answers (1)

Limey
Limey

Reputation: 12461

How about

lapply(
  seq(1, ncol(mtcars), 2),
  function(x) {
    png(paste0("histogram_", x, ".png"))
    par(mfrow=c(1, 2))
    hist(mtcars[, x])
    if (x < ncol(mtcars)) hist(mtcars[, (x+1)])
    dev.off()
  }
)

which runs the function on pairs of columns...

Here is a variation that will work for any number of rows and columns and any data frame (that contains only numeric columns).

makePlots <- function(df, nr, nc) {
  k <- nr * nc
  lapply(
    seq(1, ncol(df), k),
    function(x, df, nr, nc) {
      png(paste0("histogram_", x, ".png"))
      par(mfrow=c(nr, nc))
      for (z in x:(x+k-1)) {
        if (z <= ncol(df)) hist(df[, z])
      }
      dev.off()
    },
    nr = nr,
    nc = nc,
    df = df
  )
}

makePlots(mtcars, 3, 2)

Upvotes: 1

Related Questions