Specifications and Limits
This topic contains information about supported relational database management systems, object specifications, and source column data types in OLE DB.
Supported Relational Database Products
Microsoft® SQL Server™ 2000 Analysis Services supports the following relational database products as data sources:
- SQL Server versions 6.5 and later
- Microsoft Access 97 and later
- Oracle versions 7.3 and 8.0
Note Text columns in Oracle tables should be variable length column type when using the OLE DB Provider for Oracle. Use the OLE DB for Provider for ODBC if Oracle tables contain text columns of fixed length column type.
Specifications
The following specifications apply to Analysis Services.
| Item | Specification |
|---|---|
| Dimensions in a database | 65,535 maximum, regardless of the number of cubes or whether dimensions are shared or private |
| Levels in a database | 65,535 maximum |
| Cubes in a virtual cube | 64 maximum |
| Measures in a cube | 1,024 maximum |
| Measures in a virtual cube | 2,048 maximum |
| Dimensions in a cube | 128 maximum |
| Levels in a cube | 256 maximum |
| Levels in a dimension | 64 maximum |
| Members in a dimension | 2^31-1 = 2,147,483,647 maximum |
| Members in a parent | 64,000 maximum |
| Calculated members (server defined) in a cube | 65,535 maximum |
| Calculated members in a parent measure in session context | 31,743 maximum |
| Calculated members in a parent measure in query context | 31,743 maximum |
| Calculated members in a parent dimension member in session context | 759 maximum |
| Calculated members in a parent dimension member in query context | 759 maximum |
| Aggregations per partition | 65,535 maximum |
| Cells returned by a query | 2^31-1 = 2,147,483,647 cells maximum
Although cubes can be larger than this limit, a query that requests more than 2^31-1 cells from a cube will fail. |
| Record size for source database table | 64 kilobytes (KB) maximum |
| Length of object name (except dimension name) | 50 characters maximum when using Analysis Manager
24 characters maximum when using PivotTable® Service |
| Length of dimension name | 24 characters maximum |
| Length of aggregation prefix | 50 characters maximum |
| Maximum number of distinct states in a data mining model attribute column | 255, after which the column becomes MODEL_EXISTENCE_ONLY
For more information, see Data Mining Columns |
Note The term character in this documentation refers to a UNICODE character.
Source Column Data Types
The data type of the source column for a measure must be numeric except when the Count or Distinct Count aggregate function is used.
The data type of the source column for a dimension level must be string or numeric (except currency).
OLE DB Data Types
Columns of the following OLE DB data types may be used as measures or dimension levels. Columns types that are marked with an asterisk are supported data types for data mining model columns.
- DBTYPE_BOOL*
- DBTYPE_I1
- DBTYPE_I2
- DBTYPE_I4*
- DBTYPE_UI1
- DBTYPE_UI2
- DBTYPE_UI4
- DBTYPE_I8
- DBTYPE_UI8
- DBTYPE_R4
- DBTYPE_R8* (Note: Data mining model column inputs of this type will be cast to a 4 byte floating point number)
- DBTYPE_DECIMAL
- DBTYPE_NUMERIC
- DBTYPE_VARNUMERIC
- DBTYPE_CY
- DBTYPE_DATE*
- DBTYPE_DBDATE
- DBTYPE_DBTIME
- DBTYPE_DBTIMESTAMP
Columns of the following OLE DB data types may be used as measures only if the Count or Distinct Count aggregate function is used. These data types may be used as dimension levels.
- DBTYPE_BYTES
- DBTYPE_HCHAPTER*
- DBTYPE_STR
- DBTYPE_WSTR*
External Limitations
Limitations imposed by other technologies, such as the RDBMS being used, may limit some features of Analysis Services. For example, when merging two partitions containing a large number (> 100) of aggregations, you may receive an error message indicating that the maximum number of ODBC Access 97 File Sharing lock counts has been exceeded. This number is controlled by the Access 97 MaxLocksPerFile registry entry, not by any configuration parameter in Analysis Services.
Other such external limitations may apply as well.