Openbravo Issue Tracking System - Modules
View Issue Details
IDProjectCategoryView StatusDate SubmittedLast Update
0039387ModulesMulti-Dimensional Profit And Losspublic2018-10-02 09:442020-01-20 14:06
ReporterJONHM 
Assigned ToTriage Omni OMS 
PriorityhighSeveritymajorReproducibilityalways
StatusnewResolutionopen 
PlatformOS5OS Version
Product Version 
Target VersionFixed in Version 
Merge Request Status
Regression date
Regression introduced by commit
Regression level
Review Assigned To
Support ticket
OBNetwork customer
Regression introduced in release
Summary0039387: Performance issue on multi-dimensional profit and loss report
DescriptionPerformance issue on multi-dimensional profit and loss report. When running report for 4 months of data, the process takes too long and the server gets overloaded.
Steps To ReproduceRun multi-dimensional profit and loss report

*Org: *
*General Ledger: A general ledger with lots of data
*Year: Current year
*Period from: Feb
*Period to: May
*Column: Organization, Business Partner
*Row: Product
Proposed Solution
Additional Information
TagsNo tags attached.
Relationships
Attached Files
Issue History
Date ModifiedUsernameFieldChange
2018-10-02 09:44JONHMNew Issue
2018-10-02 09:44JONHMAssigned To => Triage Finance
2018-10-02 09:44JONHMResolution time => 1540245600
2018-10-02 11:28martinsdanIssue Monitored: martinsdan
2018-10-23 09:37JONHMResolution time1540245600 => 1541458800
2018-10-25 10:03JONHMSteps to Reproduce Updatedbug_revision_view_page.php?rev_id=17842#r17842
2018-10-25 10:12dmiguelezNote Added: 0107563
2018-10-25 10:12dmiguelezTypedefect => design defect
2020-01-20 14:06ngarciaResolution time1541458800 =>

Notes
(0107563)
dmiguelez   
2018-10-25 10:12   
This is not an Issue.

Based on the volumes of data of the client, it is impossible to do the CrossJoin they are proposing.

70 organizations * 35.048 bpartners * 31.975 products * 5.298.609 records in the fact table = 4,15 * 10^17

Mondrian itself is raising this problem:
'ResourceLimitExceededException: Mondrian Error: Size of CrossJoin result (21,974,101) exceeds limit (5,000,000)'

In this case the filters must be used to reduce the amount of data calculated and shown.

It would be nice if the application itself raises this problem to the user before attempting to launch the query.

The design defect that can be solved is to count the amount of records that each dimension is going to cross and, if it is higher than 100, raise an error forcing the user to select a filter.