I am newbie on Docker. Seems the following pipe prints nothing. If I am right, the -a stdin attaches my host stdin to the container process's stdin.

~ $ echo test | docker run -a stdin -a stdout -a stderr --rm alpine cat -

When I add the -i option, it does print the "test":

~ $ echo test | docker run -i -a stdin -a stdout -a stderr --rm alpine cat -
test

Why must we append the -i option?

I did read this post Confused about Docker -t option to Allocate a pseudo-TTY, but still confused by options like -a, -d, -t, -i. For example, is there any difference between the following two runs?

~ $ docker run -a stdin -a stdout -it ubuntu /bin/bash

~ $ docker run -it ubuntu /bin/bash

---

After some search, I think the -i option's purpose is to make the stdin of PID 1 open which by default is closed.