Very good response, and yes, for an ideal scenario, you were off by about 3. The other factor of 3 comes from the other factors that go into making a practical, usable installation, not just theory. First, the roof will not be covered, corner to corner with panels, all inclined to the latitude. Either you angle them and space them a bit to prevent them from shading eachother, or you sacrifice efficiency. In reality, you do the former. Second, you need access and room for all the other associated equipment. Next, you have to factor in that that average daily insolation is nowhere near steady throughout the year. You easily need 50% over capacity to charge batteries on less light days to reduce grid draws in cloudy days, but in truth, if you are actually trying to produce all the power on site, you need 2-3x the rated capacity. This will get you through most 3-5 day rainy periods. You will still need a grid tie for those exceptional times when it stays rainy for 2 weeks, etc, but those are rare, so nobody installs for that.
There is a reason net metering makes solar work. If you know anyone who is truly off-grid on solar+batteries (one of my best friends is), then you will find that annual averages, rated output, and other "typical" specifications don't cover real world, must have available juice 24/7 365 types of situations.
Regardless, even taking your ideal scenario, an average Walmart store could conceivably produce enough juice to recharge the bulk of the truck volume for that store. For the distribution center, not even close. A large distribution center could see that many trucks in a busy hour, much less a day.
FWIW, check out
Walmart Corporate - Photos of Walmart Trucks and Logistics Team. They have some pics in there that look strangely similar to something groundbreaking from last week. I was shocked too.