A couple of problems/confusions:
- Datasheets sometimes give "keep out" areas. This just means they don't want copper (or maybe silkscreen either) there, as there may be projections from the component which could cause short circuits or wear. A common confusion (I say "common" because I've done this a few times myself, long ago!) is to use an EDA tool's keep-out tools for this. But a keep-out object may not keep out just the offending material: in the past, Altium had one common keep-out layer, against which all layers collided. Nowadays, you can set objects by layer, but remember to set every layer you need (probably just top copper, and maybe silk; or multilayer where it really does matter, such as in-board antennas).
- Whether you assign the same pad number/name to a common feature is up to you. Keeping them distinct is most general, but clutters up the schematic -- you need a symbol with as many pins as the footprint has. (A multi-pin DFN MOSFET for example ends up looking like a hedgehog this way.) Making them common is... semantically true enough, but kind of in the opposite way you'd like the EDA tool to understand it. That is, you make them the same name so they get the same net, which means you must wire them together on the PCB. If you're going to wire them anyway (to get rated ampacity, to reduce EMI, etc.), done and done. If you only need one connection, or want to make a through connection (jumper), though -- that's different. Note that a jumper is really a net split, so should be drawn as such on the schematic -- but you can only do that by splitting up the pins, so you should do that. Or maybe you only need/want to connect one of the pins (OR instead of AND connectivity of the pads in question), such as for a tactile switch where the pins are in pairs.
Tim