My take on this is that what is important here is that the tool does not fail at this point, so we should either:
- check before the "alter" is done (and lock the table name in some way for using it at this stage), or
- try 10 times with underscores, and if this fails, generate a quasi-unique random name (try as many times as needed)
Hi Frank!
My take on this is that what is important here is that the tool does not fail at this point, so we should either:
- check before the "alter" is done (and lock the table name in some way for using it at this stage), or
- try 10 times with underscores, and if this fails, generate a quasi-unique random name (try as many times as needed)