There is one issue with this question. But let me clear one thing, even though they said max should be initialized with 0 or -1, that was with respect to Marks dataset. In the same lecture or the next lecture, they said that picking a value for initializing max or min should be based on the context of the data set.
So for marks: 0, -1 can be used to initialize Max since the possible range is 0 - 100 / 150 (whatever may be the upper case)
But for Sale, we can initialize MAX with 1 too, because will there be any bill with amount 0 (which means nothing is bought). The possible range of bills is from 1 to the upper limit. So choosing 1 is not an issue. (neglecting the case of 0.50 rs bills, that too impossible nowadays - not sure any supermarket sells anything which costs less than 1).
For more clarification, consider a scenario where we are trying to find out the max temperature in a hill station. Can we initialize MAX with -1 or 0? Absolutely not. Because temperature can have negative values, so we should pick something well below the possible range, like -100.
So we shouldn't always pick -1 or 0 to initialize Max. Should always consider the context of the data.
The problem with this question is that, If they start by initializing MAX as 1, then what if there are no particular bills for the given scenario. Even though there are no records, MAX will be zero, so should we infer that there is a record with 1 rs as the total amount or no record at all.