Grep matches only of multiple separated strings

Question

I have a file with lines containing this format:

fieldA=value1, fieldB=value2, fieldC=value3, fieldD=value4, fieldE=value5

I am interested in fieldA, fieldB, fieldD. However, fieldC may or may not be present, therefore I cannot use something like:

grep "field" * | awk -F"," '{print $1, $2, $4}'

My end goal is to have output like this, all in one line:

fieldA=value1, fieldB=value2, fieldD=value4

I tried using grep -E, but it outputs those fields in different lines, and the association between the fields breaks.

grep -o -E "field1_=\w*|field2_=\w*|field3_=\w*"

are those 3 fields' name (A,B and D) fixed? or everything is dynamic? — Kent
– Kent, Commented Jan 16, 2013 at 0:07
Kent, all the fields are fixed (A, B,..D), but the values are not. — Alisa Vladimirovna Shinkorenko
– Alisa Vladimirovna Shinkorenko, Commented Jan 16, 2013 at 1:30

Kent · Accepted Answer · 2013-01-16 00:20:56Z

1

if you know the field name of A,B,D grep and xargs could do the job. ( awk/sed could do it for sure)

grep -Po "fieldA=[^,]*|fieldB=[^,]*|fieldD=[^,]*" file|xargs -n3

that gives you:

fieldA=value1 fieldB=value2 fieldD=value4

if you want the comma in output:

grep -Po "fieldA=[^,]*,|fieldB=[^,]*,|fieldD=[^,]*" file|xargs -n3

answered Jan 16, 2013 at 0:20

Kent

196k36 gold badges248 silver badges316 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Beta · Accepted Answer · 2013-01-16 00:07:43Z

0

Is a sed solution acceptable?

sed 's/^\([^ ]* [^ ]*\).*\(fieldD=[^,]*\).*/\1 \2/' filename

answered Jan 16, 2013 at 0:07

Beta

99.4k18 gold badges159 silver badges156 bronze badges

Thank you this helped. I was able to come up with a similar solution with sed.