初次尝试ColumnStoreIndex

初次尝试ColumnStore IndexPosted on 1、首先使用非聚集索引 TIME ON SELECT MemberId , ((VExpd) (DD, (DD, FactOrders o WITH ( INDEX (IX_IsBigOrder_O) ) IsBigOrder IsBigOrder MemberId 结果 SQL Server parse and compile time:CPU time ms.SQL Serve

初次尝试ColumnStore Index Posted on

1、首先使用非聚集索引

TIME ON SELECT MemberId , ((VExpd) (DD, (DD, FactOrders o WITH( INDEX(IX_IsBigOrder_O) ) IsBigOrder IsBigOrder MemberId

结果

SQL Server parse and compile time: CPU time ms. SQL Server Execution Times: CPU time ms. SQL Server Execution Times: CPU time ms. (1121211 row(s) affected) . Scan ahead reads 57581, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0. . Scan ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0. . Scan ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0. SQL Server Execution Times: CPU time ms.

2、使用ColumnStore Index

TIME ON SELECT MemberId , ((VExpd) (DD, (DD, FactOrders o WITH( INDEX(IX_CS) ) IsBigOrder IsBigOrder MemberId

结果

SQL Server Execution Times: CPU time ms. SQL Server parse and compile time: CPU time ms. SQL Server Execution Times: CPU time ms. SQL Server Execution Times: CPU time ms. (1121211 row(s) affected) . Scan ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0. . Scan ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0. SQL Server Execution Times: CPU time ms.

从结果的logical reads上可以看出明显差异，空间，CPU time和elapsed time分别2倍和3倍之多。

3、使用ColumnStore Index不能像使用其它非聚集索引那样没有，使用ColumnStoreIndex的如下:

1.一个表只能有一个ColumnStore Index

2.不能使用过滤索引

3.索引必须是partition-aligned

4.被索引的表变成只读表

5.被索引的列不能是计算列

6.不能使用Include关键字

由于第4点的，以前怕麻烦，所以比较抗拒，从这次的尝试来看，空间，性能提升还是很大的。

计划运行一阵，跟踪瞧瞧如何。