I agree with you about having to depressurize an entire hold if you are moving cargo in total deep space. In that case your containers would either be space-rated (making them much more expensive - especially if you were putting even low-G grav plating in them), or they'd need some other form of vacuum-proofing. I got around that myself with by coming up with a cheap polymer lining that is a one-use thing. The insides of a standard container are sprayed with a thin lining making it a vacuum-sealed environment. Once the doors are closed it's sealed, and when you open it you have to reapply it. It's meant to be a cheap way way to bridge the gap, and obviously not being reusable without re-applying means you can't 'cheat' and avoid the cost of space-rated containers. It's not like PC's try to min-max everything....
Cargo airlocks do make sense, but it's also an added expense and added tonnage to a ship. Small ships can't afford that (like a free trader), but a 10,000 dton ship could do so. Smaller ships might cheat their way through this by having multiple holds, each one sealed, and then cycle from the one that has the external hatch to the ones on the inside. It works, but not efficiently.
Container types should follow the existing model like you mentioned - you have the standard ones and then you have the seagoing ones. A sea going container is built of steel and is designed to be weather resistant and much stronger to protect the internal cargo from the harsh sea environment. They are also more expensive than the cheaper aluminum or even fabric-sided containers that only need to keep the weather out. The reason for that boils down to cost. No merchant is going to pay for things they don't absolutely need - that maxim has been in force for millenia and would continue onwards into the future (now, change the race and you may change that rule). One thing on re-use though - it varies based on the amount of cargo going back to the exporting country. The US is a net-importer of goods in containers and you will see many Asia-bound ships leave empty or with small container loads (as opposed to showing up with 10-15k TEUs worth of loaded ones). The reason for that, again, is cost. Sometimes its just cheaper to make a new one in China than for that container to make its' way back from being unloaded. Seems stupid, but it's just economics. Once the price/availability of steel goes back up it becomes more economical to ship them back empty. In space trade I don't see any empty containers going back for reuse (at least on starships). The economic model there doesn't work well with running empty (even though operational costs are basically fixed if at 0% or 100% full). I read somewhere of the 16(ish) million cargo containers out there, only about 6 million are actively used to transport sea cargo - the rest are repurposed to all kinds of other things - including regular ground transport. Which since they are steel makes perfect sense.
I worked my way through college at UPS loading the sem-trailers you see on the road. For manual loading of so many different packages we actually did take advantage of a LOT of space - but except for the training film where every box was of universal size and shape (which was NOT the reality!), "efficiently" loading the truck was building walls of packages that didn't shift or crush the ones underneath. We certainly weren't tall enough to reach the top, so basically the top 18-24" was left for loading bags of small packages (tossing them up there to be more precise). Speed was valued above all else because we had timelines we had to meet for loading the packages and to give time for them to get to the railyards or for drivers to pick them up to drive to their next hub destination. Since we never had to build loads for zero-G environments I can't say how well we might be able to adapt (or not) to it using the same principle that boxes are rarely going to magically fit together as seen on the training video.
Good stuff!