CODEWITHSUNDEEP
bash
CuriousDude
CuriousDude

Reputation: 1117

Bash script, print filenames that contain a string

I have a folder with a couple of files that I need to organize/manipulate depending on if they both exist, or only one of them exists.

In my folder called folder1/checkthese/*.bam the files are:

file1_aln.bam
file1_aln_sorted.bam

I have a script that checks if I have the unsorted file (which is just *_aln.bam) and sorted file (*_aln_sorted.bam) but I am having trouble getting my script to run correctly depending on if they both exist or not.

Here is my mini script:

for files in folder1/checkthese/*.bam 
do
    if [[ ${files} =~ "_aln.bam" ]] && [[ ${files} =~ "_aln_sorted.bam" ]]
    then                                                                                                                                                                                                         
          echo "both files exist, need to delete unsorted file only"
          echo "REMOVE $(basename ${files/_aln*}_aln.bam)"
          rm -f ${files/_aln*}_aln.bam            
    elif [[ ${files} =~ "_aln_sorted.bam" ]] && [[ ! ${files} =~ "_aln.bam" ]]                                                        
    then                                                                   
          echo "Only sorted file exists, all good"                             
    fi 
done

But this is the output I get:

Only sorted file exists, all good.

But clearly the unsorted file exists so for some reason it is skipping the first part of my loop and not removing the _aln.bam file. I am not sure how to change my conditional statement in my elif statement so that if ONLY the _aln_sorted.bam file exists, then all is good and I don't need to delete anything. I think I should not be using the && for my elif statement, but I thought the ! essentially is the NOT boolean for this.

Upvotes: 0

Views: 502

Answers (3)

liborm
liborm

Reputation: 2724

I will present a little less conventional solution, stressing two points:

  • prefer working with file lists like with other textual data
  • separate the logic and destructive operations (you can check what you're about to delete)

First create some test files

mkdir data
seq 1 5 | xargs -I{} touch 'data/file_{}_aln.bam'

# first three of them have their sorted equivalents
seq 1 3 | xargs -I{} touch 'data/file_{}_aln_sorted.bam'

First let's check what files I'd delete:

find data -name '*.bam' | sort | sed 's/_sorted//' | uniq -d

The complement are the files I have to sort yet:

find data -name '*.bam' | sort | sed 's/_sorted//' | uniq -u

After checking, I can do something like this to delete the files

find data -name '*.bam' | sort | sed 's/_sorted//' | uniq -d | xargs rm

The final check if all unsorted are gone can be done easily by

ls data/*_aln.bam 

# or to get some numeric results:
ls data/*_aln.bam | wc -l

Of course the usual caveats apply - use sensible file names or you have to use find -print0 | xargs -0 and deal with the consequences.

Upvotes: 0

Arda Örkin
Arda Örkin

Reputation: 151

Dude, your comparision can't do what you want.

Your first comparision is checking for the files that name contains both _aln.bam and _aln_sorted.bam string! And the second is checking for the files that name contains _aln_sorted.bam and doesn't contain _aln.bam!

So these comparions works on same file in every execution!

You need this:

#!/bin/bash

for file in /full_path/folder1/checkthese/*.bam 
do
    if [[ ${file} =~ "_aln.bam" ]]
    then                                                                                                                                                                                                         
          echo "Unsorted file was found! It will be removed!"
          echo "Removing the file named ${file}"
          rm -f ${file}
      echo "File removed!"
    elif [[ ${file} =~ "_aln_sorted.bam" ]]                                                        
    then                                                                   
     echo "${file} is a sorted file!"
    fi 
done

Upvotes: 1

CuriousDude
CuriousDude

Reputation: 1117

-----------EDIT--------------------

Okay I fixed my original script which did not use booleans to check for strings in the filename but instead checked if files existed. This worked for me:

Originally I had this script as well but ran into similar problems:

for files in folder1/checkthese/*.bam 
do
    if [ -f ${files/_aln*}_aln.bam ] && [ -f ${files/_aln*}_aln_sorted.bam ]
    then                                                                                                                                                                                                         
          echo "both files exist, need to delete unsorted file only"
          echo "REMOVE $(basename ${files/_aln*}_aln.bam)"
          rm -f ${files/_aln*}_aln.bam            
    elif [ -f ${files/_aln*}_aln_sorted.bam ] && [ ! -f ${files/_aln*}_aln_sorted.bam ]                                                    
    then                                                                   
          echo "Only sorted file exists, all good"                             
    fi 
done

Output works now.

Upvotes: 0

Related Questions