CODEWITHSUNDEEP
sed
Loca Toney
Loca Toney

Reputation: 45

Substitute characters between specific strings

I would like to substitute every commas happen to be between two specific strings (DP & MQ) to semicolons.

input

0,0,0,DP=1,1,1,1,MQ=2,2,2

expected output

0,0,0,DP=1;1;1;1;MQ=2,2,2

I have variable number of fields before and after the DP & MQ so I thought sed would be the best. I don't want to substitute commas before DP or after MQ. Could any of you please help me I know that it should looks like this

sed 's/DP=.,.,.,.,MQ/DP=somethingMQ/g'

Thanks in advance

Upvotes: 0

Views: 143

Answers (4)

Vijay
Vijay

Reputation: 67211

below code will do:

awk -F"=" '{OFS="=";gsub(",",";",$2)}1'

tested:

> echo "0,0,0,DP=1,1,1,1,MQ=2,2,2" | awk -F"=" '{OFS="=";gsub(",",";",$2)}1'
0,0,0,DP=1;1;1;1;MQ=2,2,2

or you can use:

perl -plne '$_=~/DP=(.*)MQ/;$a=$1;$a=~s/,/;/g;$_=~s/(.*DP=).*(MQ.*$)/$1$a$2/g'

Tested:

> echo "0,0,0,DP=1,1,1,1,MQ=2,2,2" | perl -plne '$_=~/DP=(.*)MQ/;$a=$1;$a=~s/,/;/g;$_=~s/(.*DP=).*(MQ.*$)/$1$a$2/g'
0,0,0,DP=1;1;1;1;MQ=2,2,2

or 

perl -F"=" -ane '$F[1]=~s/,/;/g;print join "=",@F'

tested:

> echo "0,0,0,DP=1,1,1,1,MQ=2,2,2" | perl -F"=" -ane '$F[1]=~s/,/;/g;print join "=",@F'
0,0,0,DP=1;1;1;1;MQ=2,2,2

Upvotes: 0

Mirage
Mirage

Reputation: 31548

With awk , you can do like this(provided there are no more = there)

awk -F"=" '{gsub(",",";",$2); $1 = $1; print $1"="$2"="$3}' temp.txt

output

0,0,0,DP=1;1;1;1;MQ=2,2,2

Upvotes: 0

potong
potong

Reputation: 58351

This might work for you (GNU sed):

sed -r 's/DP.*MQ/\n&\n/;h;y/,/;/;G;s/.*\n(.*)\n.*\n(.*)\n.*\n/\2\1/' file

This sed idiom, marks the string in question (using newlines), copies the marked line, alters the string and then combines the original line with the altered string.

The marking of the string may have to be more specific i.e.:

sed -r 's/DP=[^=]*MQ=/\n&\n/;h;y/,/;/;G;s/.*\n(.*)\n.*\n(.*)\n.*\n/\2\1/' file

If only some of file may contain the string in question use:

sed -r '/DP=[^=]*MQ=/{s//\n&\n/;h;y/,/;/;G;s/.*\n(.*)\n.*\n(.*)\n.*\n/\2\1/}' file

Upvotes: 2

Kent
Kent

Reputation: 195029

if you have gnu sed: this should work with your example:

sed -r 's/(.*DP=)(.*)(MQ=.*)/echo -n \1;echo -n \2 \|tr "," ";"; echo -n \3/ge' input

test with your example

kent$  sed -r 's/(.*DP=)(.*)(MQ=.*)/echo -n \1;echo -n \2 \|tr "," ";"; echo -n \3/ge' <<<"0,0,0,DP=1,1,1,1,MQ=2,2,2"
0,0,0,DP=1;1;1;1;MQ=2,2,2

Upvotes: 0

Related Questions