How to use `jq` in a shell pipeline?

I can't seem to get jq to behave "normally" in a shell pipeline. For example:

$ curl -s https://api.github.com/users/octocat/repos | jq | cat

results in jq simply printing out its help text*. The same thing happens if I try to redirect jq's output to a file:

$ curl -s https://api.github.com/users/octocat/repos | jq > /tmp/stuff.json

Is jq deliberately bailing out if it determines that it's not being run from a tty? How can I prevent this behavior so that I can use jq in a pipeline?

Edit: it looks like this is no longer an issue in recent versions of jq. I have jq-1.6 now and the examples above work as expected.


* (I realize this example contains a useless use of cat; it's for illustration purposes only)


Solution 1:

You need to supply a filter as an argument. To pass the JSON through unmodified other than the pretty printing jq provides by default, use the identity filter .:

curl -s https://api.github.com/users/octocat/repos | jq '.' | cat

Solution 2:

One use case I have found myself doing frequently as well is "How do I construct JSON data to supply into other shell commands, for example curl?" The way I do this is by using the --null-input/-n option:

Don’t read any input at all! Instead, the filter is run once using null as the input. This is useful when using jq as a simple calculator or to construct JSON data from scratch.

And an example passing it into curl:

jq -n '{key: "value"}' | curl -d @- \
  --url 'https://some.url.com' \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json'