An $n$-by-$n$ square is regular if the two conditions are met:
- Each of the integers from $0$ to $n^2 − 1$ appears in exactly one cell, and each cell contains only one integer (so that the square is filled), and
- If we express the entries in base-$n$ form, each base-$n$ digit occurs exactly once in the units’ position, and exactly once in the $n$’s position.
So if S is a regular square (as defined above), and T is obtained from S by exchanging two rows or two columns, then is T also a regular square? Why is it a regular square?
As you have stated the definition, you can derive item 2 from item 1. Swapping columns or rows doesn't change the set of numbers that are in the cells, so it is still a regular square.
I believe you want item 2 to require that each row and column contain each base-$n$ digit once in the ones place and once in the $n$'s place.
Even so, swapping a pair of rows (or columns) will make $T$ a regular square if $S$ is. If $S$'s rows are proper, so are $T$'s, as they are the same. Each column of $T$ has the same set of entries as the corresponding column of $S$, so it satisfies 2 as well.