Bash change filenames from month name to month number

Question

I have been trying to do this all afternoon and cannot figure out how to do this. I'm running MXLinux and from the commandline am trying (unsucessfully) to batch edit a bunch of filenames (I've about 500 so don't want to do this by hand) from:

2020-August-15.pdf
2021-October-15.pdf

To:

2020-08-15.pdf
2021-10-15.pdf

I cannot find anything that does this (in a way I understand) so am wondering. Is this possible or am I to do this by hand?

Admittedly I'm not very good with Bash but I can use sed, awk, rename, date, etc. I just can't seem to find a way to combine them to rename my files.

I cannot find anything on here that has been of any help in doing this.

Many thanks.

EDIT:

I'm looking for a way to combine commands and ideally not have to overtly for-loop through the files and the months. What I mean is I would prefer, and was trying to, pipe ls into a command combination to convert as specified above. Sorry for the confusion.

EDIT 2:

Thank you to everyone who came up with answers, and for you patience with my lack of ability. I don't think I'm qualified to make a decision as to the best answer however have settled, for my use-case on the following:

declare -A months=( [January]=01 [February]=02 [March]=03 [April]=04 [May]=05\
[June]=06 [July]=07 [August]=08 [September]=09 [October]=10 [November]=11 [December]=12 )

for oldname in 202[01]-[A-za-z]*-15.pdf
do
    IFS=-. read y m d ext <<< "${oldname}"
    mv "$oldname" "$y-${months[$m]}-$d.$ext"
done

I think this offer the best flexibility. I would have liked the date command but don't know how to not have the file extension hard coded. I was unaware of the read command or that you could use patterns in the for-loop.

I have learned a lot from this thread so again thank you all. Really my solution is a cross of most of the solutions below as I've taken from them all.

Answer 1

With just Bash built-ins, try

months=(\
    January February March April May June \
    July August September October November December)

for file in ./*; do
     dst=$file
    for ((i=1; i<=${#months[@]}; ++i)); do
        ((i<10)) && i=0$i
        dst=${dst//${months[$i]}/$i}
    done
    mv -- "$file" "$dst"
done

This builds up an array of month names, and loops over it to find the correct substitution.

The line ((i<10)) && i=0$i adds zero padding for single-digit month numbers; remove it if that's undesired.

As an aside, you should basically never use ls in scripts.

The explicit loop could be avoided if you had a command which already knows how to rename files, but this implements that command. If you want to save it in a file, replace the hard-coded ./* with "$@", add a #!/bin/bash shebang up top, save it as monthrenamer somewhere in your PATH, and chmod a+x monthrenamer. Then you can run it like

monthrenamer ./*

to rename all the files in the current directory without an explicit loop, or a more restricted wildcard argument to only select a smaller number of files, like

monthrenamer /path/to/files/2020*.pdf

You could run date twelve times to populate the array, but it's not like hard-coding the month names is going to be a problem. We don't expect them to change (and calling twelve subprocesses at startup just to avoid that seems quite excessive in this context).

As an aside, probably try to fix the process which creates these files to produce machine-readable file names. It's fairly obvious to a human, too, that 2021-07 refers to the month of July, whereas going the other way is always cumbersome (you will need to work around it in every tool or piece of code which wants to order the files by name).

Answer 2

Assuming you have the GNU version of date(1), you could use date -d to map the month names to numbers:

for f in *.pdf; do 
  IFS=- read y m d <<<"${f%.pdf}"
  mv "$f" "$(date -d "$m $d, $y" +%F.pdf)"
done

I doubt it's any more efficient than your sed -e 's/January/01/' -e 's/February/02/' etc, but it does feel less tedious to type. :)

Explanation:

1. Loop over the .pdf files, setting f to each filename in turn.

2. The read line is best explained right to left:
   a. "${f%.pdf}" expands to the filename without the .pdf part, e.g. "2020-August-15".
   b. <<< turns that value into a here-string, which is a mechanism for feeding a string as standard input to some command. Essentially, x <<<y does the same thing as echo y | x, with the important difference that the x command is run in the current shell instead of a subshell, so it can have side effects like setting variables.
   c. read is a shell builtin that by default reads a single line of input and assigns it to one or more shell variables.
   d. IFS is a parameter that tells the shell how to split lines up into words. Here we're setting it – only for the duration of the read command – to -. That tells read to split the line it reads on hyphens instead of whitespace; IFS=- read y m d <<<"2020-August-15" assigns "2020" to y, "August" to m, and "15" to d.

3. The GNU version of date(1) has a -d parameter that tells it to display another date instead of the current one. It accepts a number of different formats itself, sadly not including "yyyy-Mon-dd", which is why I had to split the filename up with read. But it does accept "Mon dd, yyyy", so that's what I pass to it. +%F.pdf tells it that when it prints the date back out it should do so ISO-style as "yyyy-mm-dd", and append ".pdf" to the result. ("%F" is short for "%Y-%m-%d"; I could also have used -I instead of +anything and moved the .pdf outside the command expansion.)
   f. The call to date is wrapped in $(...) to capture its output, and that result is used as the second parameter to mv to rename the files.

Answer 3

If you are using GNU utilities and the Perl version of rename (not the util-linux version), you can build a one-liner quite easily:

rename "$(
  seq -w 1 12 |
  LC_ALL=C xargs -I@ date -d 1970-@-01 +'s/^(\d{4}-)%B(-\d{2}\.pdf)$/$1%m$2/;'
)" *.pdf

You can shorten if you don't care about safety (or legibility)... :-)

rename "$(seq -f%.f/1 12|date -f- +'s/%B/%m/;')" *.pdf

Answer 4

Another way with POSIX shell:

# Iterate over pattern that will exclude already renamed pdf files
for file in [0-9][0-9][0-9][0-9]-[^0-9]*.pdf
do
  # Remove echo if result match expectations
  echo mv -- "$file" "$(
    # Set field separator to - or . to split filename components
    IFS=-.
    # Transfer filename components into arguments using IFS
    set -- $file
    # Format numeric date string
    date --date "$3 $2 $1" '+%Y-%m-%d.pdf'
  )"
done

Answer 5

What I mean is I would prefer, and was trying to, pipe ls into a command combination to convert as specified above.

Well, you may need to implement that command combination then. Here’s one consisting of a single “command” and in pure Bash without external processes. Pipe your ls output into that and, once satisfied with the output, remove the final echo…

#!/bin/bash

declare -Ar MONTHS=(
  [January]=01
  [February]=02
  [March]=03
  [April]=04
  [May]=05
  [June]=06
  [July]=07
  [August]=08
  [September]=09
  [October]=10
  [November]=11
  [December]=12)

while IFS= read -r path; do
  IFS=- read -ra segments <<<"$path"
  segments[-2]="${MONTHS["${segments[-2]}"]}"
  IFS=- new_path="${segments[*]}"
  echo mv "$path" "$new_path"
done

Answer 6

What is working for me in Mac OS 12.5 with GNU bash, version 3.2.57(1)-release (arm64-apple-darwin21) is the following :

for f in *.pdf; do  mv "$f" "$(echo $f |sed -e 's/Jan/-01-/gi' -e 's/Feb/-02-/gi'  -e 's/Mar/-03-/gi'  -e 's/Apr/-04-/gi'  -e 's/May/-05-/gi'  -e 's/jun/-06-/gi'  -e 's/Jul/-07-/gi'  -e 's/Aug/-08-/gi'  -e 's/Sep/-09-/gi'  -e 's/Oct/-10-/gi' -e 's/Nov/-11-/gi' -e 's/Dec/-12-/gi'  )"; done

Note the original file had the month expressed in three litters in my case :

./04351XXX73435-2021Mar08-2021Apr08.pdf