For the first subtask, we can brute force.

Time Complexity: ~\mathcal{O}(NQ)~

For the second subtask, we can observe a few things. Firstly, notice that for every update event, only a row and a column is updated. Thus, we can keep track of the rows and columns separately. Let ~r_j~ represent the number of times row ~j~ was updated, and let ~c_i~ represent the number of times column ~i~ was updated. For an update event, we only need to increment ~r_j~ and ~c_i~ by one. For a query event, we simply add ~r_j + c_i~, and modulo by ~2~.

However, notice that position ~(i, j)~ is counted twice in both the row and the column. Thus, we need to use a hashmap or a map to keep track of the number times position ~(i, j)~ was the actual position that was updated, and subtract that value.

Time Complexity: ~\mathcal{O}(Q)~ or ~\mathcal{O}(Q \log Q)~