DA Morgan wrote:
I do not know SQL Server's Cluster Index, can't comment there.
DB2 does not have an "index only table".
There are two table types that have "some" overlap:
* Range Clustered Tables (RCT). These are essentially arrays on disk.
Typically used for for static look up tables
* Multi dimensional clustering. (MDC)
Organizing the data in an n-dimensional cube.
The individual blcoks are indexed instead of the rows and located by
intersecting (or or-ing) the dimensions. As a result block indexes are
tiny with one entry potentially covering thousands of rows.
So in a way both these approaches are "index-less" tables. The end
result is the same: Fast access without duplication of data.
Cheers
Serge
--
Serge Rielau
DB2 Solutions Development
IBM Toronto Lab
Doesn't DB2 have a table or index type comparable to Oracle's
Index Organized Table or SQL Server's Cluster Index where that
would not be the case?
Index Organized Table or SQL Server's Cluster Index where that
would not be the case?
DB2 does not have an "index only table".
There are two table types that have "some" overlap:
* Range Clustered Tables (RCT). These are essentially arrays on disk.
Typically used for for static look up tables
* Multi dimensional clustering. (MDC)
Organizing the data in an n-dimensional cube.
The individual blcoks are indexed instead of the rows and located by
intersecting (or or-ing) the dimensions. As a result block indexes are
tiny with one entry potentially covering thousands of rows.
So in a way both these approaches are "index-less" tables. The end
result is the same: Fast access without duplication of data.
Cheers
Serge
--
Serge Rielau
DB2 Solutions Development
IBM Toronto Lab