The SPARSE command tells Analytic Services to use the sparse data extraction
method, which optimizes performance when a high proportion of the reported
data rows are #MISSING
. This data extraction method is different
from the regular method. Analytic Services cannot use the sparse data retrieval
optimization method on Dynamic Calc or Dynamic Calc And Store members.
<SPARSE
If you have at least one sparse row dimension in your report, Analytic Services uses the sparse data extraction method in two cases:
Case 1: You use SUPMISSINGROWS in your report script to suppress #MISSING
values, and Analytic Services estimates that a very high proportion
of the requested data rows are #MISSING
. In this case,
Analytic Services implicitly uses the sparse method to optimize performance.
Case 2: You explicitly use the SPARSE command in your report script. This forces Analytic Services to use the sparse method. If you use the SPARSE command in a report, and you have not used SUPMISSINGROWS, Analytic Services automatically turns on SUPMISSINGROWS for the report containing SPARSE. Analytic Services also turns on SUPMISSINGROWS for all following reports in your report script, unless you specify INCMISSINGROWS in a subsequent report.
Note: If your report does not contain at least one sparse row dimension, Analytic Services cannot use the sparse method, and reverts to the regular method. Analytic Services displays a message to tell you that it cannot use the sparse method.
When Analytic Services uses the sparse method, it displays the following message: "Report Writer Sparse Extractor method will be executed."
If you have at least one sparse row dimension in your report, the report
is very large, and a very high proportion of the reported data rows are
#MISSING
, you may want to use the SPARSE command to force Analytic Services
to use the sparse data extraction method. You can then assess if
this improves your report script performance.
If your report requests a small number of cells (#MISSING
and
non-missing), the sparse data extraction method is less efficient than
the regular method. In this case, Analytic Services uses the regular method,
unless you have at least one sparse row dimension in your report, and you
explicitly use the SPARSE command.
SPARSE method: When Analytic Services uses the sparse data extraction method, Analytic Services first selects the row member combinations you have requested in your report script. Analytic Services looks at only the non-missing data blocks for these row member combinations. If your database is very sparse, this method is very efficient.
Regular method: By contrast, when Analytic Services uses the regular
data extraction method, it cycles through every possible member combination
requested by the report script. It then reports only those rows that are
not #MISSING
.
For example, suppose that only 1 in 10,000 data cells exist in a database.
The remaining cells are #MISSING
. On this database, you run a
report script that requests 100% of the data,
and uses SUPMISSINGROWS to
suppress the #MISSING
values.
If Analytic Services uses the regular method of data extraction, it cycles through all the requested member combinations.
If Analytic Services uses the sparse extraction method, it looks only at the non-missing data blocks for the row member combinations requested. As this database is very sparse (only 1 in 10,000 data cells exist), the number of existing data blocks is probably low. The sparse method produces the report much faster.
Note: The sparse extraction method cannot be used if the report contains attribute dimensions.
©2004 Hyperion Solutions Corporation. All Rights Reserved. http://www.hyperion.com |