multi-line grep not making list - grep

Question

I have a grep line that someone else provided for me that I don't know how to change.
The original line was this:
grep id=\"desc\"* $ADDON_SETTINGS | awk -v ORS=, '{gsub(/"/, "");print $2}' | tr -s 'value=' ' ' | sed 's/ //g' | echo "[$(cat)]"
And it pulled from a file that contained the following (this is a sample segment):
<settings>
<setting id="cfirst" value="false" />
<setting id="cicons" value="false" />
<setting id="days" value="3" />
<setting id="delay" value="0.000000" />
<setting id="desc01" value="10" />
<setting id="desc02" value="18" />
<setting id="desc03" value="6" />
<setting id="desc04" value="13" />
<setting id="desc05" value="6" />
...
It pulled out the value for lines with "desc" in the id and resulted in a list:
10,18,6,13,6...
Now the program that generates the data file has changed the data to look like this:
<settings version="2">
<setting id="allc" default="true">false</setting>
<setting id="cfirst" default="true">false</setting>
<setting id="cicons" default="true">false</setting>
<setting id="days">3</setting>
<setting id="delay" default="true">0</setting>
<setting id="desc01">10</setting>
<setting id="desc02">18</setting>
<setting id="desc03">6</setting>
<setting id="desc04">13</setting>
...
I figured this might be easier as I just need to pull the value between > and < but if I use this:
grep id=\"desc\"* $ADDON_SETTINGS | awk -v ORS=, '{">|<";print $3}' | echo "[$(cat)]"
But it's not working right. Not sure what I'm missing.

Answer 1

Try:
$ awk -F'[<>]' '/"desc/{printf "%s%s",c,$3; c=","} END{print""}' file
10,18,6,13
How it works:
-F'[<>]'
This tells awk to use < or > as field separators.
/"desc/{printf "%s%s",c,$3; c=","}
For any line that containst "desc, this tells awk to print the variable c followed by the third field. The third field is the number we want. The variable c is initially the empty string but after the first print, we set it to a comma, ,. This causes the numbers we want to be printed, each separated by a comma.
END{print""}
After we have finished reading the file, this tells awk to print a newline character.

Answer 2

The reason your solution doesn't work is '{">|<";print $3}', which doesn't make sense. The expression ">|<" doesn't do anything. You want a line like:
$ grep id=\"desc input.txt | awk -F"<|>" '{print $3}'
However, a single awk solution is:
awk 'match($0,/id=\"desc[0-9]+\">([0-9]+)/, a){printf "%s%s",sep,a[1];sep=","} END{print ""}' input.txt
10,18,6,13
or: with a file:
$ cat tst.awk
match($0,/id=\"desc[0-9]+\">([0-9]+)/, a){
printf "%s%s",sep,a[1];sep=","
}
END{print ""}
$ awk -f tst.awk input.txt
10,18,6,13
Explanation:
the match with regex id=\"desc[0-9]+\">([0-9]+) puts the number between parentheses in a[1].
Print a[1] with the separator sep, which hasn't a value the first time around.
END: you need to print a newline.

Answer 3

Your content has html/xml format. The proper way would be to use html/xml parsers.
xmlstarlet solution:
Sample input.html content:
<settings version="2">
<setting id="allc" default="true">false</setting>
<setting id="cfirst" default="true">false</setting>
<setting id="cicons" default="true">false</setting>
<setting id="days">3</setting>
<setting id="delay" default="true">0</setting>
<setting id="desc01">10</setting>
<setting id="desc02">18</setting>
<setting id="desc03">6</setting>
<setting id="desc04">13</setting>
</settings>
The job:
res=($(xmlstarlet sel -t -v "//setting[contains(#id, 'desc')]" 1.html | tr '\n' ' '))
This will extract <setting> tag values with attribute id containing "desc" and make them items of the array res
Check the 2nd array item value:
echo ${res[1]}
18

Answer 4

grep digits between desc\d+"> and <
grep -oP 'desc\d+">\K\d+(?=<)' file | paste -sd ","
This will capture digits between desc\d+"> and <.
Note:desc\d+ would resolve to desc01, desc02 etc
-o is used to capture groups
-P tells it is a perl regex
\K is a lookbehind assertion which tells to start capturing data
(?=) is a lookahead assertion which here tells to stop capturing data once < is found