An indexed view serves a different purpose than a table. Also - if there is no table, then there can be no view.
A table is for storing actual data.
A view can be used for presenting data from tables in a user-friendly manner, for instance by replacing foreign key values that are substitute keys, with the natural key or a full name from the lookup table, or by adding user-friendly aliases for column names.
A view can be used to hide sensitive or irrelevant information for specific users. You can grant users permissions on the view, and not on the table, so the hidden information is truly inaccessible to them.
A view can be used to 'pre-process' joins, making life easier when you create queries - you don't have to set up the joins between the underlying tables again and again.
A view is the only object in SQL Server where you can create a single index that covers columns from more than one table. This is useful for queries where you have criteria spread over multiple tables.
I don't think there is much point in adding an index on a view if it covers a (set of) column(s) from only a single underlying table: such an index should already be defined on the table.
转载于:https://www.cnblogs.com/programmingsnail/archive/2011/03/17/1986915.html
本文探讨了SQL中视图与表的区别,视图如何提供更友好的数据展示,隐藏敏感信息,简化查询创建,以及如何利用视图为跨表查询预处理数据。同时,讨论了在视图上创建索引的有效性和适用场景。
3396

被折叠的 条评论
为什么被折叠?



