How to Insert Millions Records into Database Table

Suppose you need to insert millions of rows to database table. There are many ways that can accomplish this task fast, very fast or whatever. But the question is to perform the insert and not to produce locks on selects while the system is online and running which is more complicated. It should work the following way:

1. create "proxy table"
2. make batch load into it: either BCP or bulk insert or DTS
3. move from "proxy table" into production via "insert into - select from", using partitioning (if there are millions of rows), to prevent locks on selects, for example additional identity field in "proxy table" would make partitioning easy.

Related Posts:

• T-SQL: Update and Output in a Single Query
• MSSQL 2005 beta 2 DE first impressions
• MSSQL Overusability
• SQL Server 2005 Developer Edition Beta 2 (English)
• MSSQL 6.5 Service Pack 5 is Available
• SQL Server 2005 Documentation and Samples
• SQL Express is not ready yet
• Scheduling algorithm using MSSQL sysjobschedules
• Sql Server 2005 EE issues
• Moving SQL