Character replacement of Textfile

[syintel87]

I'd like to replace "Mh:####:Mh" by "Mh" in a text file,
where #### ranges from 0000 to 3451.

Would you give me some tips how to use sed command to achieve this goal?

For example, when I ran one command, sed 's/Mh/m/' Mh.txt > Mh_tr.txt, it worked.
However, not only when there are combination of special characters with numbers, but also when there exists a numerical variable, I have no idea about this.

Thank you in advance.

[jimmybee]

Sorry but its not really clear what you want to search and replace. Can you give us an example of a line you want to achieve?

[syintel87]

Sorry but its not really clear what you want to search and replace. Can you give us an example of a line you want to achieve?

Sorry for making you confused.

My files look like the following:
FA file >Mh:0000:MhA1_Contig0AGTTGTTAATTTTTC
GFF file Mh:0000:MhA1_Contig0 Freeze3 mRNA 4027 5970 + ID=MhA1

My goal is to make these files look like the following:
FA file >MhA1_Contig0AGTTGTTAATTTTTC
GFF file MhA1_Contig0 Freeze3 mRNA 4027 5970 + ID=MhA1

In other words, I will have to
- replace Mh:0000:Mh by Mh
- or delete Mh:0000:
- or delete :0000:Mh

Regardless of what numbers are in between Mh's, only "Mh" should be remained. In other words, the pattern Mh:####:Mh, where #### varies from 0000 to 3451, has to be reduced to "Mh".

I hope my description does make sense.
Thank you in advance for sparing your precious time.

[jimmybee]

Looks like your way is ok but you need to escape the colons. This might not be the most elegant way of doing it (im far from a expert in sed and awk) but

sed 's/Mh\:.*\://' file > new_file

Basically it searches for Mh:####: and removes it to leave the second "Mh"

Hope that helps

[BAMseek]

Just in case there is another ":" on that line, and you don't want to be too greedy, you might try the more conservative

sed 's/Mh\:[0-3][0-4][0-5][0-1]\://' file

[syintel87]

Dear jummyybee and BAMseek,

Thank you so much!!!