Grep Ternary Curl Reponse

Question

I'm trying to return a boolean if a curl response has a status of 200.

curl https://www.example.com -I | grep ' 200 ' ? echo '1' : echo '0';

This however brings back:

grep: ?: No such file or directory
grep: echo: No such file or directory
grep: 1: No such file or directory
grep: :: No such file or directory
grep: echo: No such file or directory
grep: 0: No such file or directory
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:--  0:00:01 --:--:--     0

So I'm thinking (and from reading other threads) that grep doesn't support the ternary operator?

Is there a simple one liner to do this without creating a shell script? (also my grep ' 200 ' is loose but I figure I can make that more specific later)

Answer 1

Like a quick study of the grep manual page should reveal, it allows for options, a search pattern (multiple patterns if you specify multiple -e options), and an optional list of file names, falling back to reading standard input like many other Unix filters if no file names are specified. There is nothing about a ternary operator.

If it were supported, it would arguably need to be a feature of the shell, but no shell I know of supports anything like this. It would also clash with the syntax for ? as a wildcard metacharacter which (in isolation) matches any single-character file name.

The usual shorthand idiom in Bourne-compatible shells looks like

curl https://www.example.com/ -I | grep -q ' 200 ' && echo 1 || echo 0

This could accidentally match on a "200" somewhere else in the output. For improved accuracy, try curl https://www.example.com/ -I | head -1 | grep ' 200 ' or perhaps curl https://www.example.com/ -I | awk 'NR==1 { print ($2=="200" ? "1" : "0"); exit }' where the precision of the matching operation is significantly improved.

As an afterthought, if you want to print "0" in the case of invalid URLs and other unexpected errors too, you could set the default output to 0 and change it to 1 when you see 200, then print whatever it ended up as at the end.

curl https://www.example.com/ -I |
awk 'BEGIN { s=0 } NR==1 { s=($2=="200"); exit } END { print s }'

Answer 2

All the other answers are discussing your proposed solution, bit I would like to suggest that this is an XY problem. What you are probably looking for is a reliable, programmatic way to check whether curl succeeded or not. For that, you should just add -f:

curl -fqI http://www.example.com

This exits with code 0 on success and non-zero on any kind of failure (including server errors that successfully complete an HTTP session but report an HTTP error). In a shell script, for example, you can access this exit code with $? right after the command if you want to store it in a variable, or use the shell's if/&&/|| to conditionally execute other code:

if curl -fqI http://www.example.com
then echo 1
else echo 0
fi

If the goal really is to just print success/not success, most shells have an option to automatically display nonzero exit codes; for example, in zsh with setopt printexitvalue in your ~/.zshrc, you will see something like this (curl output suppressed to highlight the interesting bit):

% curl -fqI http://www.example.com >/dev/null 2>&1 
% curl -fqI http://www.aoeuaoeuaoeuaoeu.com >/dev/null 2>&1
zsh: exit 6     curl -fqI http://www.aoeuaoeuaoeuaoeu.com > /dev/null 2>&1

Answer 3

! curl https://www.example.com -I | grep -q ' 200 '; echo "$?"
! curl https://www.example.com -I | grep -q ' 200 '; yes "$?" | sed q

Answer 4

You can use awk to process the output as:

curl https://www.example.com -I | awk '/ 200 /{found=1} END{if (found) print 1; else print 0}'

This sets the variable found to 1 (although anything would work in this case) when the pattern / 200 / is seen. Then when the processing is done, the if tests to see if found has been set and reacts accordingly.

Answer 5

Here's how I might do it:

if [ "x$(curl -Isqw '%{http_code}' -o /dev/null http://example.com/)" = 'x200' ]; then echo "yep"; else echo "nope"; fi

To break it down:

1. Set the "silent" and "quiet" flags on curl and redirect the actual response to the bit bucket with -o /dev/null. This stops all curl output whatsoever, which lets us focus on our quarry, the status code. (The -I is a courtesy to the server: we are discarding the response anyway, so we request only the headers of the response)
2. Ask curl to output the response code with the -w '%{http_code}'flag.
3. Catch all the output in a single string with the "$()"syntax, and prepend the letter x just to make sure there's always something in the string, even if the command crashes and burns (old habits die hard). The proper thing to do would be to check the return code of the curl command at this point, but since any failures now get neatly mapped to a "not 200" result anyway, we won't bother.
4. Use the [ command (usually found at /bin/[, which is often a symlink to /bin/test) to check if the resulting string is identical to "x200"
5. Now we already have the boolean value, but since the question asked for a "string boolean", finally use if to choose what to output. The actual syntax for if may vary from one shell to another; this one will work with sh, bash and their relatives.