I have some csv like below
uid_01, joe, yes, no uid_02, sam, yes, maybe uid_03, c, ruth, yes, maybe uid_04, **alan**, no, yes I want to replace alan with ruby, uid will be unique here. I need to change depending on $position, any idea how to get that in sed?
Also I want to give both pattern and replacement value as arguments.
Tried something like this, but working only at last entry
sed -i '/^'$uid',/s/[^,]*$/'$returnvalue'/' $OUTPUT Thanks in advance, please help!
awk suits it better:
awk -v id='uid_04' -v repl='ruby' 'BEGIN {FS=OFS=", "} $1 == id {$2 = repl} 1' file.csv uid_01, joe, yes, no uid_02, sam, yes, maybe uid_03, c, ruth, yes, maybe uid_04, ruby, no, yes 
gnu-awk then use awk -i inplace -v id='uid_04' -v repl='ruby' 'BEGIN {FS=OFS=", "} $1 == id {$2 = repl} 1' file.csv
gawk -i inplace, perl -i, ruby -i, sed -i, etc.) takes a copy of the file so - do you really care about taking a copy of the file or do you just not want to do it explicitly in your code with cmd file > tmp && mv tmp file and would prefer cmd -i file or cmd -i inplace file?This might work for you (GNU sed):
id='uid_04' name='ruby' position=2 sed -i "/^$id,/s/[^,]*/ $name/$position" file If the id is uid_04 replace the second field by ruby
