


Ninja is a boy who lives in ninjaland. Every day, during the morning he gets 1 coin from his mother. He wants to buy exactly ‘N’ candies. Each of the candies cost 2 coins usually and 1 coin if it is on sale. Ninja has to buy exactly K[i] candies of the i-th type(he buys candies in the evening).
Ninjas can buy any(possibly zero) number of candies of any type during any day(if he has enough money to do it). If the candy is on sale he can buy it for 1 coin and otherwise he has to buy it for 2 coins.
There are ‘M’ special offers in the candy shop. The j-th offer (D[j], C[j]) means that candies of the C[j]-th type are on sale during the D[j]-th day.
Ninja wants to buy all candies as soon as possible. Your task is to calculate the minimum day when he can buy all the candies.
The first line of input contains an integer ‘T’ denoting the number of test cases to run. Then the test cases follow.
The first line of each test case contains two space separated integers ‘N’ and ‘M’ denoting the total number of types of candies and the number of special offers in the candy shop.
The second line of each test case contains ‘N’ space-separated integers K[i], denoting the number of candies of each type Ninja has to order.
The next M-lines contain two space-separated integers (D[i], C[i]), denoting that the candy of C[i] type is on sale on the D[i]-th day.
Output Format :
For each query, print the minimum day when the Ninja can order all candies.
Output for each test case will be printed in a new line.
Note:
You do not need to print anything; it has already been taken care of. Just implement the given function.
1 <= T <= 10
1 <= N, M <= 10^4
0 <= K[i] <= 10^4
1 <= C[i] <= N
1 <= D[i] <= 10^4
Here N denotes the total number of candies and M denotes the total number of special offers.
Here K[i] denotes the number of candies of the i-th type Ninjas has to order.
Here C[i] and D[i] denote that the candy if the C[i]-th type will be on sale on the D[i]-th day.
Sum of K[i]’s in each test case will be less than 5000.
2
4 4
1 0 2 0
1 1
1 3
2 1
2 3
3 3
1 1 1
1 1
1 2
1 3
4
5
For test case 1, Ninja can buy candies in the following manner-
• Buy candy of type 1 on day 1 for 1 coin.
• Buy candy of type 3 on day 2 for 1 coin.
• Buy candy of type 3 on day 4 for 2 coins.
For test case 2, Ninja can buy candies in the following manner-
• Buy candy of type 1 on day 1 for 1 coin.
• Buy candy of type 2 on day 3 for 2 coins.
• Buy candy of type 3 on day 5 for 2 coins.
1
5 6
1 2 0 2 0
2 4
3 3
1 5
1 2
1 5
2 3
8
Can we iterate over days to get our answer?
We will iterate over all possible days, which will be between 1 to 2*(sum of all K[i]). Let our current day be ‘currDay’. We will check if any valid distribution is possible for the current day.
If we had several days for some type of candies, let's use the last one, it is always not worse than some of the previous days. Then let's iterate over all days from 1 to ‘ansd’ and do the following:
The algorithm will be-
O(sumCandy * sumCandy), where sumCandy is the total number of candy which Ninja has to buy.
The time complexity due to the linear search will be O(sumCandy). For each check, we are iterating over all days. Thus, the total time complexity will be O(sumCandy * sumCandy).
O(sumCandy * sumCandy), where sumCandy is the total number of candy which Ninja has to buy.
The space complexity for each check will be O(sumCandy) due to storing all the jobs which have the i-th day as the last day of sale. As we performing this check for sumCandy times in the worst case, the overall space complexity will be O(sumCandy*sumCandy).