创建具有一对多关系的表或仅创建具有单列的表

Question

If I have multiple key value pairs in Azure Blob Storage such as:

  -/files/key1
  -/files/key2
  -/files/key3

And each key is uploaded by a user, but a user can upload multiple keys, what is the best table design in my SQL database to reference what keys are associated with what user?

A) Table with single column - Everytime I add a file to BLOB storage I add a row to a single column table with the username and the associated key value ie:

 AssociationColumn
 -User1+key1
 -User2+key2
 -User1+key3

Will this be slow in looking up all the keys for User1 for example if I query using some sort of regex starts with? Will making this two column with User as one column and key as another column affect performance at all? How can I achieve this one to many relationship?

Also is it bad to store keys using an identifier such as 1-2-n? Any suggestions on how to create unique identifiers that can fit in the space of varchar(MAX)?

Answer 1

The correct approach in a relational database is to have a junction table. This would have at least two columns:

• User
• Key

You wouldn't put these in a single column -- I don't think even in Azure.