How to use array members as the search or replace pattern in bash?

Question

I have about a thousand PDF files with names like Foo_Jan_2013.pdf, Bar_Feb_2012.pdf, Foo_Mar_2013.pdf, Bar_Mar_2013.pdf. I want to rename these files replacing the month names with numbers, such as Foo_01_2013.pdf, Bar_02_2012.pdf, Foo_03_2013.pdf, Bar_03_2013.pdf.

One approach I tried is to define two arrays (one for the match pattern and other for the replace pattern), then use the respective array members in the search-replace, like so:

match=(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec);
replace=(01 02 03 04 05 06 07 08 09 10 11 12);

for FILE in *.pdf;
do 
    for i in {0..11};    
    do
        echo ${match[i]} ${replace[i]};  # Note, see below
        echo ${FILE/${match[i]}/${replace[i]}};
    done
done

The line marked Note prints Jan 01, Feb 02, etc. as expected, but using them as the search or replace pattern in the next line has no effect, and $FILE gets printed as it is.

I have also confirmed that both of the following work:

${FILE/Jan/01};

foo=Jan;
bar=01;
${FILE/$foo/$bar};

Is it possible to use an array member as the pattern, as I attempted above? If so, what is the correct syntax? If not, what other options do I have to solve this problem?

What version of bash are you using? Check $BASH_VERSION. Your script works on my Mac running 3.2.48(1)-release and my Linux running 3.2.25(1)-release. — rob mayoff
– rob mayoff, Commented Oct 5, 2013 at 17:06
Instead of echo can you try: rename -n "s/${match[i]}/${replace[i]}" — devnull
– devnull, Commented Oct 5, 2013 at 17:11
Works for me in bash 4.2.45(0) as well. Doesn't seem like the best way though; I'll suggest another alternative as an answer. — torek
– torek, Commented Oct 5, 2013 at 20:27
@tripleee That's right! The main reason I am doing this rename is to allow the files to be sorted easily. I didn't mention it here because it wasn't relevant to the problem. — Masked Man
– Masked Man, Commented Oct 6, 2013 at 6:14

iscfrc · Accepted Answer · 2013-10-05 17:23:37Z

2

BASH supports associative arrays which would be a good fit for your use case:

# Initialize the associative array
declare -A months

# Fill in the array with the month string as the key and the zero-padded month number as the value
num=0
for m in Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec; do
  ((num++))
  months[$m]=$(printf '%02d' $num)
done

# Test your examples
for file in Foo_Jan_2013.pdf Bar_Feb_2012.pdf Foo_Mar_2013.pdf Bar_Mar_2013.pdf; do
  month=$(cut -d_ -f2 <<<"$file")
  new_name="${file/$month/${months[$month]}}"
  echo "$file -> $new_name"
done

Output:

Foo_Jan_2013.pdf -> Foo_01_2013.pdf
Bar_Feb_2012.pdf -> Bar_02_2012.pdf
Foo_Mar_2013.pdf -> Foo_03_2013.pdf
Bar_Mar_2013.pdf -> Bar_03_2013.pdf

answered Oct 5, 2013 at 17:23

iscfrc

4312 silver badges6 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

chepner Over a year ago

It's worth pointing out that associative arrays were added in bash 4.0.

torek · Accepted Answer · 2013-10-05 20:50:14Z

2

For situations like this my favorite trick is to generate a series of commands with (e.g.) sed:

num=0
subst=
for month in Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec; do
    num=$(($num + 1))
    case $num in ?) num2=0$num;; *) num2=$num; esac
    subst="$subst -e 's/\(.*\)_${month}_\(.*\)/mv & \1_${num2}_\2/'"
done
# echo "sed $subst" # debug
# echo Foo_Jan_2013.pdf | eval sed $subst # debug
ls *.pdf | eval sed $subst # debug
# ls *.pdf | eval sed $subst | sh # once it all works

(this version avoids bash-specific code). I left in the debug-echo steps, too, and left the final part commented-out.

answered Oct 5, 2013 at 20:50

torek

499k71 gold badges764 silver badges891 bronze badges

Comments

Community · Accepted Answer · 2017-05-23 12:29:32Z

1

I put together the answers by iscfrc, torek, and the comment by devnull, and solved the problem as follows:

# Initialize an associative array with (month name, number) pairs
declare -A map=([Jan]=01 [Feb]=02 [Mar]=03 [Apr]=04 [May]=05 [Jun]=06 
[Jul]=07 [Aug]=08 [Sep]=09 [Oct]=10 [Nov]=11 [Dec]=12);

# For each month name, use 'rename' command on filenames having the month name
for m in ${!map[@]};
do
    eval rename \'s/$m/${map[$m]}/\' *$m*.pdf
done

With this approach, each filename is matched only against the pattern that it does contain (not all 12 month names), and the nested loop structure in my question is avoided. It also doesn't use any additional pipes, and works even when the "separator" is different from _.

edited May 23, 2017 at 12:29

CommunityBot

11 silver badge

answered Oct 6, 2013 at 5:56

Masked Man

1

Collectives™ on Stack Overflow

How to use array members as the search or replace pattern in bash?

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related