remove all occurences of dot only between two strings and not in the rest of the file

Question

apologies if this has been asked before - I could not find an answer to my question though

assuming there is the following file

touch test2

echo "refH.fasta = ref/GCA_013924565.1_ASM1392456v1_genomic.fasta" >> test2


echo "subjectGCA_017717955.1_PDT000990484.1_genomic_querry.fasta = GCA_017717955.1_PDT000990484.1_genomic.fasta" >> test2

echo "file=subjectGCA_017717955.1_PDT000990484.1_genomic_querry" >> test2

in the above file I would like to remove the dots ONLY between the strings 'subject' and '_querry' but not in the rest of the file.

Therefore the output should look like this:

refH.fasta = ref/GCA_013924565.1_ASM1392456v1_genomic.fasta
subjectGCA_0177179551_PDT0009904841_genomic_querry.fasta = GCA_017717955.1_PDT000990484.1_genomic.fasta
file=subjectGCA_0177179551_PDT0009904841_genomic_querry

thanks

Answer 1

Here is a Ruby to do that:

ruby -lpe '$_=$_.split(/(subject.*?_query)/).
    map{|s| s=s[/subject.*?_query/] ? s.gsub(/\./,"") : s}.join' test2

Or Perl:

perl -lnE '@a=(); for $x (split /(subject.*?_query)/){ 
    $x=~s/\.//g if $x=~/subject.*?_query/; 
    push @a,$x } 
    say join("",@a)' test2

It is possible entirely in Bash:

while IFS= read -r line || [[ -n $line ]]; do 
    if [[ $line =~ (subject.*_query) ]]; then 
        line=${line/""${BASH_REMATCH[1]}""/""${BASH_REMATCH[1]//./}""}
    fi     
    printf "%s\n" "$line"
done <test2

But this only handles one match per line.