filesize - R Saving similar objects of very different sizes results in identical file sizes

Question

I'm having an issue with saving dist-class objects (with saveRDS) and getting much larger file sizes than expected. The code to reproduce this problem can be found at https://github.com/kstagaman/stackoverflow-help

---

Below reads in a phyloseq object with a very large community matrix (47 x 1,060,929)

library(phyloseq)

obj.size <- function(x) {paste("Object size:", format(object.size(x), units = "auto"))}
f.size <- function(x) {paste("File size:", round(file.size(x) / 1024^2, 1), "Mb")}

ps <- readRDS("example_phyloseq.rds")

Next, I create a distance matrix to determine community dissimilarities between samples It should result in a 47 x 47 matrix

dist.mat <- phyloseq::distance(ps, method = "bray") # creating a distance matrix
attributes(dist.mat)$Size

## [1] 47

Okay, as expected, but…

cat(obj.size(dist.mat), sep = "
")

## Object size: 453.3 Mb

This is oddly big for a distance matrix of this size (like orders of magnitude larger)

After investigating I realize that this is because attributes(dist.mat)$call is huge, because it includes every single value from the original community matrix.

dist.mat.small <- dist.mat
attributes(dist.mat.small)$call <- NULL
cat(obj.size(dist.mat.small), sep = "
")

## Object size: 12.5 Kb

Okay, so that’s looking good. Much closer to what I’d expect for the size. But…when I try to save…

saveRDS(dist.mat, file = "distance_matrix_original.rds")
saveRDS(dist.mat.small, file = "distance_matrix_small.rds")

cat(f.size("distance_matrix_original.rds"), sep = "
")

## File size: 112.5 Mb

cat(f.size("distance_matrix_small.rds"), sep = "
")

## File size: 112.5 Mb

What’s happening here? Why are the file sizes the same for objects of very different size?

Answer 1

等待大神答复