| T O P I C R E V I E W |
| dbooth |
Posted - 04/24/2012 : 03:05:17
I have a cube of sales orders with appropriate hierarchies in place, however the users like to specify some filters and then see the underlying data. This is fine but the cube data is spooled over the network, resulting in IE pages which are massive - 700,000k to 1.4g being typical. They then export the result to Excel/CSV to get back a 250k output file.
Is it possible to entirely host the cube on the server and just deliver the page (or a cached set of pages) to the user, and if they then want to export the result, build it and send only the resulting file over the network?
Regards
Dave |
| 4 L A T E S T R E P L I E S (Newest First) |
| admin |
Posted - 04/24/2012 : 14:28:17
Thanks for reporting the problem. We will see if there is a way to optimize the performance of paging.
Correct, a report can have multiple subscriptions and each subscription can have different filter value. |
| dbooth |
Posted - 04/24/2012 : 13:43:50
Paging is set up at 30 records a page. As the report builds, depending on the filter settings, it can compile up to 1400 pages. As the page count takes place the size of memory used by the IE process climbs rapidly. Changing the filter also shows memory used by IE rising as the page count goes up.
The OLAP settings we currently use are 30 records per page, Enable server formatting, and Server Side Execution. So 30 records x 1400 pages, 42000 records in this one example, which consumes upward of 800MB memory in IE.
The user browsing the cube wants to extract 8 results, changing one filter 8 times to get the differing datasets. So he could construct 8 reports and create a subscription for each, or I guess, 8 subscriptions with differing Filter settings? |
| admin |
Posted - 04/24/2012 : 12:14:05
Have you tried using paging? Paging would let you see one page at a time and export the rest from the server.
Currently, it is possible to retain Drillthrough selections for a user and cube. In a future release we can add the ability to save the Drillthrough selection at any level.
What do you mean by “[Drillthrough] not being available in Subscription reports”? |
| dbooth |
Posted - 04/24/2012 : 05:07:02
The fix to this was in the report design. Instead of having a report which consists of 30000 flattened rows, now it starts with the aggregation (Depot in this case) and the user then uses Drillthrough to obtain the raw data in detail. That's very fast in comparison.
Is it possible to retain Drillthrough selections not just for a Cube (which is already available), but for a specific Report? That would make Drillthrough more useful, and compensate for it not being available in Subscription reports.
Regards
Dave |
|
|
Report Portal
