My app written in C# and loads over 1 milion rows from database. For some reason, I have to perform this initial loading and then I'll search inside the loaded rows instead of querying SQL and every single column has large data( about 500 characters each row) and each column has different type . For search, I have to check the equality of data from selected record with the 3 specific columns from other records.
I think using Binary Search Tree can do it fast because now , it takes minutes using nested for loops(2 loops). If I can put the data in the correct structure, making HashCode of 3 columns will create unique integer data and put that in the BST.
I do not know any structure suits this hypothesis. The main question is, is this hypothesis correct? Or just dreamed?