如果将来表增长到千万条记录级,到时候(也许两年后),再在表上建索引,可以吗?到时候建立的索引,会自动把表中所有的数据都“索引”吗?作者: ILoveMK 时间: 2009-09-18 10:03
提示: 作者被禁止或删除 内容自动屏蔽作者: TOADLover 时间: 2009-09-18 10:28 标题: 回复 #2 ILoveMK 的帖子 呵呵!!
精辟!作者: TOADLover 时间: 2009-09-18 10:29
嗯,你这里的是译文,是哪里来的?
英文地址也可以,这段话哪里来的。。。作者: Minsic 时间: 2009-09-18 11:03 标题: 回复 #4 TOADLover 的帖子 应该来自于《oracle9i&10g编程艺术》一书……作者: ILoveMK 时间: 2009-09-18 11:17
提示: 作者被禁止或删除 内容自动屏蔽作者: TOADLover 时间: 2009-09-18 13:32
嗯,谢谢两位!
-----------------------
In order to understand wht is a clustered index in oracle it is imp. to know what is a cluster.
A cluster is a method to store data for more than 1 table in the same block.
This can be compared to a condition where we have a join condtion of 2 tables(primary key of one is the foreign key in other).
Under such conditions we need to index the cluster i.e. which data segment belongs to which table.
Such type of clustering is Indexed clusetering in oracle. 这个说的是索引簇,好像不是索引?
Plz make a note that clustered index has some other meaning in other data base esp. Sybase
--------------------------
按照这里的说法,cluster index是不可能创建在普通表上的了?
---------------------------
[ 本帖最后由 TOADLover 于 2009-9-18 13:34 编辑 ]作者: TOADLover 时间: 2009-09-18 13:37
A cluster index is unlike a table index in the following ways:
----------------------------------------------------------------------------
1. Keys that are all null have an entry in the cluster index.
2. Index entries point to the first block in the chain for a given cluster key value.
3. A cluster index contains one entry per cluster key value, rather than one entry per cluster row.
4. The absence of a table index does not affect users, but clustered data cannot be accessed unless there is a cluster index.