First, I'd try having the lens calibrated by Sigma (I don't know why they can't seem to do a better job of this, right from the factory... There seem to be an awful lot of these lenses that need adjustment!)
You will see the focus error greater and more obviously at closer distances, when depth of field is most shallow and both foreground and background blur are greatest. So it's going to be more obvious at the largest apertures, too.
Before sending it in, I'd take a bunch of test shots to check the lens and see if it's pretty consistent in it's error. Try it at different distances, various apertures, and with different subjects. Also shoot the same targets and setups as closely as possible with a different lens on your camera, to be sure it's not the camera that is out of adjustment.
You might want to print some examples of the errors or burn some sample images to a CD/DVD to send in with the lens, to show them the problem.
Next, retest the lens in the same ways when it comes back from Sigma. Hopefully you will then be happy with it.
If not, the Canon 50/1.4 is the closest alternative. It doesn't seem to have anywhere near as much calibration issues, but does have some nuances of it's own:
I think the Siggy is better wide open. It's both sharper and gives slightly nicer background bokeh. The Canon needs to be stopped down a little to give good results, it's a little soft wide open. Either "use" the softness to your advantage, or stop down a little if you want maximum possible sharpness. It gets better each third stop and the softness is usually completely gone by about f2 or f2.2. This really is nothing surprising or unusual... most lenses simply aren't at their very best wide open. The Sigma being optimized to use wide open is probably a lot more unusual.
Conversely, the Sigma seems to start getting a little soft around f5.6 or smaller apertures. I don't know why this is the case, perhaps diffraction sets in faster with it... But the Canon remains sharp to smaller apertures (f8 on crop cameras, at least f11 on full frame).
So if absolute sharpness is a major factor for you, consider how you will be using the lens most... wide open or stopped down... to help you decide if the Canon might be a better for you.
Also, the Sigma is somewhat more robust. This means it is larger and heavier, using a ridiculuously large filter for a 50mm lens, but less prone to breakage. The Canon is decent mid-grade build, but uses a somewhat unusual (for Canon) hybrid USM to focus. It's sort of a cross between the cheaper micro motor and the better ring-type USM.
The Canon 50/1.4 focuses plenty fast, is quiet and accurate. But it's focus mechanism is more easily damaged than most. Two things seem to be the main causes... One is manually overriding the AF a lot, without first turning off AF at the switch. Canon calls it a Full Time Manual USM lens, but using it that way a lot seems to more rapidly wear out the mechanism. The solution is pretty simple... Rely on AF or turn it off at the switch before doing much manual focusing.
The other thing that seems to damage the focus mechanism is a hard bump on the focusing ring, such as might happen while the lens is being toted around in a camera bag. It seems to help to get the matched lens hood and use it... and to store the lens with the hood reversed. This largely protects the focusing ring and isn't a big deal or a bad idea to do with any lens, anyway.
Some folks also think it's a good idea to set the lens to infinity before storing it. This will retract the lens barrel, which extends a little for closer focusing, and the idea is that it will be better protected at the infinity setting. Again, it's something I do with most lenses anyway, so no big deal. I have no idea if it really makes a difference, but my Canon 50/1.4 was bought used about eight years ago (I have no idea how old it really is), hasn't been given all that special treatment and still works fine. It's always been used and stored with the lens hood (I can't say about previous owner(s), but it came with the hood when I bought it). And I really don't have much reason to manually override focus.
Lensrentals.com used to publish repair statistics on their lenses, and their numbers with a bunch of Canon 50/1.4 are a bit scary, at first glance. They didn't say how many copies of the lenses they had, but saw about 20% needing repairs in the first year's use. But, also consider that rental lenses spend a lot of time bouncing around in the back of UPS vans and airplanes, though they are hopefully packed well. And of course we have no idea if the lenses were fitted with the lens hoods, nor if the people who rented them faithfully used the hoods, or manually overrode AF frequently, or even handled the lenses as carefully as they would one that they owned. Best guess... Rental lenses probably lead a fairly tough life and that kind of usage is a pretty good "torture test" for them.
Also, Lensrentals buys in large batches direct from the manufacturer or a wholesale importer/distributor... Because of that, many of any particular model of lens are likely to have been manufactured at the same time, with subcomponents from the same batches, and tested and inspected by the same guy at the end of the assembly line. So, there is always a chance that they got a "bad batch".
On the other hand, the above "quirks" of the Canon 50/1.4 are all pretty well documented elsewhere from user reports. Canon really is way overdue coming out with a Mark II version of this relatively old lens design (AFAIK, unchanged since mid 1993), with some tweaks to the optical formula and more reliable, true USM. They used to call to the 50/1.4 their "reference" lens... The one lens in the system that sets the standards in color, contrast, etc. for all other lenses in the system. (Actually, many manufacturers used to say this about their 50mm-ish "standard" lens. It's not unique to Canon.)
