search for pattern line by line from the second field and print out the first field corresponding to the line where the pattern was found

Question

I have the following 0.txt file below separating content by columns (fields):

'Disinfectants', 'Brand A', 'Brand B', 'Brand C'
'brand A', 'below brand C', 'greater than brand B'
'brand B', 'greater than brand D', 'below brand A'

I would like to find (from the second column) every time a pattern (say "brand A") occurs, and print out the contents of the first column belonging to the line where this pattern is found. For both the content of the resulting file is like:

Disinfectants
brand B

I have seen other similar issues but only printing the column itself where the pattern was found, usually using grep.

EDIT UPDATE: from @jubilatious1 suggestion , I found a question (https://stackoverflow.com/a/9153113) on the OS as part of my search for a solution.

awk '/brand A/{ print substr( $1, RLENGTH )}' 0.txt > 1.txt

but my 1.txt output has been different than expected as it prints only part of the content of the first field (column):

'brand
'brand

Besides, just using awk '/brand A/{ print substr( $1, RLENGTH )}' I can't specify that the search only works from the second field (column) onwards for each line.

EDIT UPDATE 1: maybe just fix the output of awk '/brand A/{ print substr( $1, RLENGTH )}' so that correctly printing the contents of the fields in the first column is a first step.

Answer 1

Hackish pipeline:

cut -d, -f2- 0.txt | grep -ni 'brand a' | sed 's/:.*/p/' | sed -nf- 0.txt | cut -d, -f1

• Split on commas and omit field 1
• grep for line numbers with 'brand a' (case insensitive)
• turn line numbers into {linenumber}p -- a sed command to print that line
• pipe those sed commands to sed -nf- ... this will print only when instructed to from stdin ... so you get only the lines you want
• split on commas and print only the first field

Or perl:

perl -lanF, -e 'print $F[0] if grep /brand a/i, @F[1..$#F]' 0.txt

• Autosplit into @F on commas, and print the first field if 'brand a' (case insensitive) is found in any of the other fields.

Both output this:

'Disinfectants'
'brand B'

You can strip the single quotes however you'd like, or, you can change the split regex for perl autosplit:

perl -lanF"/[',]+/" -e 'print $F[1] if grep /brand a/i, @F[2..$#F]' brand.txt

To get this:

Disinfectants
brand B

... note that once the line starts with a split delimiter, $F[0] is an empty string.