You're testing the coil, not the entire vehicle. Yes, you should have no continuity between the primary/secondary windings and the metal of the mounting tab.
That would also mean no continuity all the way to battery negative...but each inch of conductor, each connection increases resistance and makes it harder to determine whether or not there is continuity. So testing at the coil mounting tab--and leaving it at that--is the logical answer.
Kinda guessing you're getting false/flaky readings which may or may not be based on some sort of weird capacitance/coupling between your meter and the vehicle.
Or bum test leads.
Or maybe a bum meter.
When my meter gave me crazy results like that, I contemplated sending my multimeter back to Fluke.
Instead, I put different meter leads from a cheap multimeter into the sockets of the meter, and the crazy readings stopped.
That lead me to buy new, quality meter leads. Amazon has more than a dozen brands; be sure to get silicone insulation not PVC.
As for ohm readings when you're expecting a result under about 5 ohms--and certainly under one or two ohms--you'll need to measure the resistance of your test leads and subtract that amount from your meter reading. My new test leads have 0.2 ohms all on their own. If the meter shows 0.6 ohms, I'd have to mentally adjust that downward by 0.2 ohms for a true result of 0.4.
