How to move month digits after year digits for multiple filenames in a folder with R?

user_16050990 0
Viewed 53

I have a folder with multiple raster .tif files from January 1950 to December 2018. However, they are named with the month first and then the year (see below):

[1] "./WI_only_cmi60_01_1950.tif" "./WI_only_cmi60_01_1951.tif" "./WI_only_cmi60_01_1952.tif"

[4] "./WI_only_cmi60_01_1953.tif" "./WI_only_cmi60_01_1954.tif" "./WI_only_cmi60_01_1955.tif"

[7] "./WI_only_cmi60_01_1956.tif" "./WI_only_cmi60_01_1957.tif" "./WI_only_cmi60_01_1958.tif"

...


[820] "./WI_only_cmi60_12_2010.tif" "./WI_only_cmi60_12_2011.tif" "./WI_only_cmi60_12_2012.tif"

[823] "./WI_only_cmi60_12_2013.tif" "./WI_only_cmi60_12_2014.tif" "./WI_only_cmi60_12_2015.tif"

[826] "./WI_only_cmi60_12_2016.tif" "./WI_only_cmi60_12_2017.tif" "./WI_only_cmi60_12_2018.tif"

When I bring these into R and use the Raster package to stack these:

# list tif files in working directory
tifs <- list.files(pattern = ".tif$", full.names = TRUE)

# stack tifs in working directory
rstack <- stack(tifs)

They are ordered with all the January .tif files first followed by all the February .tif files etc when I need them for each year then each month (so chronologically ordered from January 1950 - December 2018).

Is there a way to rename these files where the order of the filenames can be rearranged so character 15 and 16 of each filename is moved after the year characters (18, 19, 20, 21)?

i.e. the first filename listed would change from

"./WI_only_cmi60_01_1950.tif"

to

"./WI_only_cmi60_1950_01.tif"
raster file-rename r
4 Answers

We could capture as a group and rearrange the backreference

sub("(\\d+)_(\\d+)(\\.tif)", "\\2_\\1\\3", "./WI_only_cmi60_01_1950.tif")

-ouptut

[1] "./WI_only_cmi60_1950_01.tif"
user_37322710

I would not rename the files, but instead sort the filenames appropriately. That should be a better approach in the long run for reproducibility and updating.

Example (unsorted)

ff <- c("./WI_only_cmi60_01_1950.tif","./WI_only_cmi60_01_1951.tif", "./WI_only_cmi60_01_1952.tif", 
"./WI_only_cmi60_06_1950.tif", "./WI_only_cmi60_06_1951.tif", "./WI_only_cmi60_06_1952.tif", 
"./WI_only_cmi60_12_1950.tif", "./WI_only_cmi60_12_1951.tif", "./WI_only_cmi60_12_1952.tif")

Using Akrun's expression

i <- sub("(\\d+)_(\\d+)(\\.tif)", "\\2_\\1\\3", ff)
fs <- ff[order(i)]
fs
#[1] "./WI_only_cmi60_01_1950.tif" "./WI_only_cmi60_06_1950.tif"
#[3] "./WI_only_cmi60_12_1950.tif" "./WI_only_cmi60_01_1951.tif"
#[5] "./WI_only_cmi60_06_1951.tif" "./WI_only_cmi60_12_1951.tif"
#[7] "./WI_only_cmi60_01_1952.tif" "./WI_only_cmi60_06_1952.tif"
#[9] "./WI_only_cmi60_12_1952.tif"

A more basic approach to achieve the same

x <- gsub("WI_only_cmi60_", "", basename(ff))
d <- paste(substr(x, 4, 7), substr(x, 1, 2), sep="-")
i <- order(d)
ff[i]

Given that the pattern seems to be rather simple (69 years, 12 months each) you could also do (with all your files)

i <- rep(1:69, 12)
fs <- ff[i]

(always double check the results!)

user_6352450

Using strsplit.

x <- "./WI_only_cmi60_01_1950.tif"
revfun <- function(x) {
  r <- rev(el(strsplit(x, '')))
  Reduce(paste0, rev(r[c(1:4, 10:12, 5:9, 13:length(r))]))
}
revfun(x)
# [1] "./WI_only_cmi60_1950_01.tif"
user_65740380

The code I have used is below - via the answers given from akrun and Robert Hijmans - but I wanted to clarify how I used these answers to read all of the .tif files within a working directory and stack these:

setwd("C:/...")

# list tif files in working directory
ff <- list.files(pattern = ".tif$", full.names = TRUE) 
i <- sub("(\\d+)_(\\d+)(\\.tif)", "\\2_\\1\\3", ff) 
fs <- ff[order(i)] 

library(raster)

# create stack of tif files
rstack <- stack(fs)
user_160509900
Related