Do it right: 1NF does not require duplicate values. Each element in the proposed array of 7560 elements belongs to its own row.
By putting each element on its own line, you give RDBMS the ability to do things that it cannot do otherwise, for example: calculate statistics on a set, verify that each element adheres to domain rules, calculate differences between two sets, count / select sets general characteristics .
i will end in millions of lines (possibly more than 50 million). I'm not sure if the database can handle this without performance issues.
This is not particularly much, and you do not have to deal with all 50 million in most cases. Calculate for yourself how much access is required to search the binary tree to find one entry in a billion. The answer may surprise you.
James K. Lowden
source share