I have two DataBase Table (SQL CE). A Teacher table and a A Class table. The two tables have One-to-Many relationship where one teacher has many classes (i.e. Class has a foreign key teacher_id). Number of teachers (rows) is inserted (or generated) through C# code in run time, so as classes
Which of the following is faster in INSERT and SELECT?
- Each time a new teacher is INSERTed, a new Class Table is created (e.g. Class_teacher001) to store whichever classes the teacher has. In this case, each Class Table doesn't have to be so large and foreign key is not needed because table name would identify itself. But there will be one Teacher table and many Class_xxx Tables
- Only one Teacher table and one Class table. Each class row has a foreign key pointing at the Teacher table. Only one Class table, but it will get very long. I worry searching and reading wil be slow
Regardless of which is faster, (2) is the way to go....simply create indexes to support your searches. This is how almost all relational databases are used.
The nightmare of maintaining option (1) makes me shudder
OK, where to start. First, the relationship between Teacher and Class is potentially many-to-many, but as described by you is at least one-to-many.
The first option is absolutely the wrong way to go. Never dynamically create tables. The second option is how this sort of thing is handled. Databases are powerful, written by very smart people (usually), and can handle many more rows than all the students at a given school.
As long as you properly index your tables, they can easily support hundreds of millions of records.
I also agree with Mitch Wheat. Because when you create an index your table physically sort according to our Teacher Be creating Combined Index of (Teacher_Id ,Class_Id). Though its will Help to get fast retrieval Of Select Statment.
Unless you are already having performance problems, I would not worry about them. There are many things that can cause performance problems other than the number of rows, and they should be dealt with differently depending on what they are. You have to worry more about the number of columns in a table affecting performance than you do about the number of rows. Also the number of concurrent connections to the database. One million rows in a table is not that many, it is the other two items in conjunction with that many rows that will make a database slow. You should use the second option.
来源：https://stackoverflow.com/questions/5999641/insert-many-rows-to-one-table-or-insert-rows-separately-to-many-table