Managing the Client Cache

PivotTable® Service maintains a local cache on the client computer. When PivotTable Service executes a query, the cache is used to store the data locally. If the data is used more than once, PivotTable Service does not need to request the data multiple times. The contents of this cache and when they are updated changes with the value of the Default Isolation Mode property. The following table describes this property's values.

Isolating the Client Application from External Updates

In isolation mode, the cache is populated with axis and cell data each time a query is executed. If the original source data is changed, for example, as a result of a writeback by another user, the client cache is not updated. In OLE DB, this value is referred to as Repeatable Read mode (ISOLATIONLEVEL_REPEATABLEREAD).

In nonisolation mode, the cache is populated with axis data each time a query is executed. However, the cell data for the query is not populated. The cell data is populated only when the client application itself requests it (that is, refers to it). If the cell data is never referred to, the cache never receives it. The cache may also be refreshed if Analysis Services sends a refresh notification to PivotTable Service; in this case, PivotTable Service invalidates the existing cache. If the client application requests either axis or cell data, PivotTable Service refreshes its cache from Analysis Services. In OLE DB, this value is referred to as Read Committed mode (ISOLATIONLEVEL_READCOMMITTED).

In either mode, executing a REFRESH CUBE statement refreshes the cache.

See Also

Isolation Levels

Default Isolation Mode Property