0

I am trying to rename a couple of files with the same extensions with a differing variable in the middle. I have tried the following with no luck as to what I wish to accomplish.

file: cKO_mESC_Rep_1_H3K27Ac.gc_corrected.bam

INBAM=$(ls *_*.gc_corrected.bam)
INPUT=$(echo $INBAM | sed 's/_*.gc_corrected.bam/_input.gc_corrected.bam/g')

echo $INPUT 
cKO_mESC_Rep_1_H3K27Ac_input.gc_corrected.bam

However, I want my output to be cKO_mESC_Rep_1_input.gc_corrected.bam

Anyone know what the mistake is? I have tried so many things and I may not be searching using the right keywords to get an answer. Thanks!

Improve this question
4
  • Could you please post the name of a file AND the desired filename AFTER renaming? That would make things easier. And sed is likely the wrong tool here anyway. How about rename? Commented Jan 14, 2020 at 17:50
  • file name cKO_mESC_Rep_1_H3K27Ac.gc desired name cKO_mESC_Rep_1_input.gc_corrected.bam I'm using sed in a loop so that it looks for the paired input file in another folder that has the same perceding name just different variable in front to the extension. Commented Jan 14, 2020 at 17:55
  • There are several potential bugs in your script. A better solution is the rename command, the one written by Larry Wall (apt install rename). It is sed for file-names. Commented Jan 14, 2020 at 19:16
  • in sed and reg.exp., _* means "many _". Instead of sed 's/_*.gc.... try sed 's/_.*.gc or sed 's/_[^_]*.gc... Commented Jan 15, 2020 at 8:36

1 Answer 1

Reset to default
1

Welcome to the site.

I think you are mixing up "wildcard characters" (a.k.a. "shell globs") with regular expressions in your use of sed.

Your intention clearly is to replace the pattern "underscore, followed by any number of letters and digits, followed by .gc_corrected.bam" by _input.gc_corrected.bam. Unfortunately, your sed expression is flawed in two ways:

  • Your matching part uses the shell-typical "wildcard character" notation. However, sed expects a regular expression here, and in regular expressions, the * does not mean "any string of zero or more characters", but "zero or more repetitions of the previous character", so that your expression would replace any pattern consisting of "zero or more underscores, followed by .gc_corrected.bam" with the replacement _input.gc_corrected.bam. That is why in your case, simply the last underscore before the filename suffix is replaced with _input.
  • Even so, if you have multiple underscores in your actual filename, a pattern matching "a string consisting of any number of characters" would also include underscores, which may lead to undesired behaviour on the length of the matched string. In particular, regular expressions are greedy, and if not constructed carefully, you might end up replacing the entire _mESC_Rep_1_H3K27Ac part of the filename by _input.

In your case, the correct regular expression would be:

sed 's/_[^_]+\.gc_corrected\.bam/_input.gc_corrected.bam/g'

This would replace a string, starting with an underscore _, followed by one or more characters that are not underscores ([^_]+), followed by .gc_corrected.bam, with your substitution _input.gc_corrected.bam.

Note also that in regular expressions, the . stands for "any single character" (which in shell globs would be represented by the ?), so if you want to match a literal ., you have to escape it. This is of course not necessary in the replacement string as that is not a regular expression.

Improve this answer

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.