Example 1:
Input: privileges = [["A"],["B"],["C"]], grants = [[0,1],[1,2]]
Output: [["A"],["A","B"],["A","B","C"]]
Explanation:
Example 2:
Input: privileges = [["READ"],["WRITE"],["DELETE"],["EXPORT"]], grants = [[0,2],[1,2],[2,3]]
Output: [["READ"],["WRITE"],["DELETE","READ","WRITE"],["DELETE","EXPORT","READ","WRITE"]]
Explanation:
You are given privileges, where privileges[i] is the list of privileges directly assigned to role i. There are privileges.length roles, numbered from 0 to privileges.length - 1.
You are also given grants, where grants[i] = [source, target] means role target inherits all privileges from role source.
The grant graph is a DAG. Return a list answer where answer[i] contains every unique privilege role i has after inheriting privileges from all of its ancestors. Each answer[i] must be sorted lexicographically.
Role 2 inherits from role 1, which inherits from role 0.
Role 2 inherits both READ and WRITE; role 3 inherits role 2 and keeps its own EXPORT privilege.
1 <= privileges.length <= 20000 <= privileges[i].length <= 1000 <= grants.length <= 10000grants[i].length == 20 <= source, target < privileges.lengthsource != targetgrants forms a directed acyclic graph