minimal example to show that md5sums of two files containing the same figure are different when saved as pdfs but not pngs due to the "created at" timestamp embedded in the pdf metadata

compare_figure_md5sums.R

#
# file: compare_figure_md5sums.R
#
# description: minimal example to show that md5sums of two files
# containing the same figure are different when saved as pdfs but not
# pngs due to the "created at" timestamp embedded in the pdf metadata
#
# usage: Rscript compare_figure_md5sums.R
#
# requirements: tidyverse
#
# author: jake hofman (gmail: jhofman)
#

library(tidyverse)

########################################
# GENERATE A FIGURE, SAVE AS PDF AND PNG
########################################

# this seems not to matter, but just to be safe
set.seed(42)

# generate first version of a plot

p1 <- data.frame(x = 1:5, y = 1:5) %>%
  ggplot(aes(x, y)) +
  geom_point()

# save a pdf and a png version
pdf1 <- paste(tempfile(), '.pdf', sep = '')
ggsave(p1, file = pdf1, width = 4, height = 4)
png1 <- paste(tempfile(), '.png', sep = '')
ggsave(p1, file = png1, width = 4, height = 4)


# add a delay so "created at" times differ 
Sys.sleep(2)

########################################
# DO THE SAME EXACT THING AGAIN
########################################

# this seems not to matter, but just to be safe
set.seed(42)

# generate the same plot

p2 <- data.frame(x = 1:5, y = 1:5) %>%
  ggplot(aes(x, y)) +
  geom_point()

# save a pdf and a png version
pdf2 <- paste(tempfile(), '.pdf', sep = '')
ggsave(p2, file = pdf2, width = 4, height = 4)
png2 <- paste(tempfile(), '.png', sep = '')
ggsave(p2, file = png2, width = 4, height = 4)

########################################
# COMPARE MD5 CHECKSUMS
########################################

# pdfs differ
print("### comparing pdfs")
system(sprintf('md5 %s %s', pdf1, pdf2))

# example output:
# MD5 (/var/folders/67/k9k0tm253s18k4yty1sl3k9h0000gn/T//RtmpkQzQyF/file88816dc03691.pdf) = 5944905b7051d32614cbe95e41644567
# MD5 (/var/folders/67/k9k0tm253s18k4yty1sl3k9h0000gn/T//RtmpkQzQyF/file8881612c52d8.pdf) = c314a7fb262324fba4bc8085d582882c

# pngs match
print("### comparing pngs")
system(sprintf('md5 %s %s', png1, png2))

# example output:
# MD5 (/var/folders/67/k9k0tm253s18k4yty1sl3k9h0000gn/T//RtmpkQzQyF/file888119f5ef38.png) = 9935417db872915ee86e6b3533a1c9bc
# MD5 (/var/folders/67/k9k0tm253s18k4yty1sl3k9h0000gn/T//RtmpkQzQyF/file888152d71a8c.png) = 9935417db872915ee86e6b3533a1c9bc