Getting word index by a delimiter in a variable - bash

Question

Given line, delimiter, and word I want to get the index place of that word in the line based on the delimiter. As simple/short as possible. So for:
line="this-is-a-line_with-some.txt"
delimiter="-"
word="some"
echo <code goes here>
# should come out as 4
Of course I can split it with an array, and print the first occurrence of the word with a for loop, as follows:
line="this-is-a-line_with-some.txt"
delimiter="-"
word="some"
index=0
IFS="$delimiter" read -ra ary <<<"$line"
for i in "${ary[#]}"; do
if [[ $i == ${word}* ]]; then echo $index ; break ; fi
index=$((index+1))
done
But I'm sure there is a simpler solution.

Answer 1

simpler solution.
Replace delimiter with newline and get line numbers with grep.
<<<"$line" tr "$delimiter" '\n' | grep -n "$word" | cut -d: -f1
Minus 1:
<<<"$line" tr "$delimiter" '\n' | grep -n "$word" | cut -d: -f1 | awk '{print $1 - 1}'
# shorter
<<<"$line" tr "$delimiter" '\n' | grep -n "$word" | awk -F: '{print $1-1}'
Or really anyway just awk:
<<<"$line" awk -v RS="$delimiter" -v word="$word" '$0 ~ word{print NR-1}'

Answer 2

Understanding from OP's code and/or comments:
looking for the first occurrence of a ${delimiter}-delimited field that starts with ${word}
location index is 0-based
if ${word} is not found we generate no output
OP's code can be further reduced by using the array's 0-based index (ie, eliminate the need for the index variable):
IFS="$delimiter" read -ra ary <<<"$line"
for i in "${!ary[#]}"
do
[[ "${ary[i]}" == ${word}* ]] && echo "${i}" && break
done
# line="this-is-a-line_with-some.txt"
4
# line="a-some_def-xy-some.pdf"
1
NOTE: if ${word} is not found this will generate no output
A variation on this paramater substitution solution from superuser:
newline="${line%%${word}*}" # truncate string from 1st occurrence of ${word}
if [[ "${newline}" != "{line}" ]] # if strings are different then we found ${word}
then
IFS="${delimiter}" words_before=( ${newline} ) # break remaining string by "${delimiter}" and
# store in array words_before[]
echo "${#words_before[#]}" # number of array entries == index of 1st occurrence of ${word}
fi
# line="this-is-a-line_with-some.txt"
4
# line="a-some_def-xy-some.pdf"
1
NOTE: if ${word} is not found this will generate no output
One awk idea:
awk -F"${delimiter}" -v ptn="${word}" '{for (i=1;i<=NF;i++) if (index($i,ptn) == 1) {print i-1; exit}}' <<< "${line}"
# line="this-is-a-line_with-some.txt"
4
# line="a-some_def-xy-some.pdf"
1
Or using an inline replacement for ptn/${word}:
awk -F"${delimiter}" '{for (i=1;i<=NF;i++) if ($i ~ /^'"${word}"'/) {print i-1; exit}}' <<< "${line}"
# line="this-is-a-line_with-some.txt"
4
# line="a-some_def-xy-some.pdf"
1
NOTE: if ${word} is not found these awk scripts will generate no output
To get ideas for the truly shortest piece of code OP could try posting # codegolf, though the really short answers will likely require locating/installing new software (libs and/or binaries)

Answer 3

A solution without loop or external tool :
line="$delimiter$line"; lin2="${line%$delimiter$word*}"
if test "$lin2" != "$line"; then
IFS="$delimiter" read -ra ary <<<"${lin2#$delimiter}"
echo ${#ary[#]}
fi