CODEWITHSUNDEEP
regexstringbash
Shahin
Shahin

Reputation: 1316

regex manipulation and string substitution - alternative to sed

I have the following file

cat file.txt

ID      Location
MNS1    NC_000004.12:g.d.a144120555T>C;NC_001423.23:c.a144120513G<C
MNS2    NC_000142.12:g.a144120552C,N>D
MNS3    NC_000142.12:g.a144120559C>N

I would like to replace the input in this manner:

ID      Location
MNS1    NC_000004.12:144120555;NC_001423.23:144120513
MNS2    NC_000142.12:144120552
MNS3    NC_000142.12:144120559

I would like to remove everything but numbers that appear between : and ;

For example, I tried:

echo "NC_000004.12:g.d.a144120555T>C;" | sed 's/:[^0-9]*/:/g; s/[^0-9]*;/;/g; s/[^0-9]*$//g'

DESIRED OUTPUT

NC_000004.12:144120555

Upvotes: 0

Views: 90

Answers (2)

Toto
Toto

Reputation: 91518

If Perl is an option for you:

cat file.txt 
ID      Location
MNS1    NC_000004.12:g.d.a144120555T>C;NC_001423.23:c.a144120513G<C
MNS2    NC_000142.12:g.a144120552C,N>D
MNS3    NC_000142.12:g.a144120559C>N

perl -ape 's/:\D+(\d+).*?(?=;|$)/:$1/g' file.txt 
ID      Location
MNS1    NC_000004.12:144120555;NC_001423.23:144120513
MNS2    NC_000142.12:144120552
MNS3    NC_000142.12:144120559

Explanation:

s/            # substitute
    :           # colon
    \D+         # 1 or more non digits
    (\d+)       # group 1,, 1 or more digit
    .*?         # 0 or more any character but bewline, not greedy
    (?=;|$)     # positive lookahead, make sure we have semi-colon or end of line
/             # with
    :           # colon
    $1          # content of group 1 (i.e. the digits)
/g            # end, global

Upvotes: 1

rho
rho

Reputation: 817

This might do the trick!

sed -i.bak 's/g\.//g; s/\w>\w//g' filename

for (NC.*?): concat, bit explanation about the end output will help , although this might work:

s/NC[0-9]?:/:/

Upvotes: 2

Related Questions