Considering database performance, please review this example..
If there are only 50,000 widgets in table one,two and three (total) and 50,000,000 widgets in table four.
Using the following relational/join table to handle the first three relationships:
ID (int)
IndexID (int)
RelationType (int)
RelationI开发者_开发问答D (int)
Would it be ok just to use this relational table for table four as well (by changing the relationtype to Table4), but considering it has a signifigantly larger amount of data than the first three tables? Or should table four have its own relational/join table:
ID (int)
IndexID (int)
RelationID (int)
You can have the same relational table for table4 too and create a non clustered index on relationId.
Performance in inserting might be little slow but traversal and relating will be done fast.
精彩评论