strstr() alternative for bash

Question

I need to search a string for a substring, if the substring is found print the string upto the end of the substring.i.e

str="this is a long string"
substring="long"
expected="this is a long"

I have tried bash string manipulation and failed. Tried to use an awk command, but I can’t get it right.
This works if substring is not in a variable, but I require it in a variable since the input varies.

awk -F'long' '{print $1}' <<<$str

awk -v myvar="$substring"  -F'myvar' '{print $1}' <<<$str

prints the whole string.
Any help will be appreciated.

Asked By: mwas

||

Source

Answer 1

If you looked for bash only solution, one of the options would be:

sfx="${str#*${substring}}"
expected="${str%"${sfx}"}"

Which uses prefix removal (of anything + $substring) to determine the trailing part. And then uses suffix removal to trim that from $str yielding your expected value.

Since you seem to be OK with answers beyond just shell (like using awk), here’s similar with sed:

echo "$str" | sed -e 's#^(.*'"${substring}"').*#1#'

-> match the whole line with anything up to and including $substring being saved in the first group and replace that line with content of the first matched group.

note: both of these examples are not greedy and will match up to the first $substring.

Answered By: Ondrej K.

Answer 2

Including error handling for the case that the substring is not present:

str="this is a long string"
substring="long"
if [[ $str =~ (.*"$substring") ]]
then
  expected=${BASH_REMATCH[1]}
else
  echo ERROR: Substring not found >&2
fi

While the quotes around $substring are not necessary for your concrete example, you would need them if substring contained special regex characters.

Answered By: user1934428

Answer 3

# echo "$str" | awk -F'long' '{print $1FS}'

this is a long

Answered By: mrqiao001

strstr() alternative for bash

Question:

Answers: