MantisBT - ParaView
View Issue Details
0014565ParaView(No Category)public2014-02-11 12:262014-08-29 11:00
David C. Lonie 
David C. Lonie 
normalminorhave not tried
closedfixed 
git-master 
4.24.2 
Sandia
14565-composite-queries
incorrect functionality
0014565: Python Calculator and Extract Selection broken for composite datasets
To reproduce with Find Data:
1) Open can.ex2, enable EQPS array, advance to some interesting time (frame ~20 or so).
2) Open Find Data dialog
3) Search cells for query "EQPS == max(EQPS)"

Expected: The cell(s) with the maximum EQPS value should be highlighted.

Result: The query is run independently on each block of the dataset, and the cell(s) with the block's maximum EQPS value is highlighted.

To reproduce with python filter:
1) Same as above.
2) Add a python calculator
3) Evaluate an expression such as "EQPS + max(EQPS)"

Expected: A result array shifted by the composite maximum EQPS value should be added to the dataset.

Result: Each block is shifted independently by its maximum.
No tags attached.
child of 0013794closed David C. Lonie Remote server, find max error 
Issue History
2014-02-11 12:26David C. LonieNew Issue
2014-02-11 12:26David C. LonieAssigned To => David C. Lonie
2014-02-11 12:26David C. LonieRelationship addedchild of 0013794
2014-02-11 12:42David C. LonieTopic Name => 14565-composite-queries
2014-02-11 12:42David C. LonieNote Added: 0032307
2014-02-11 12:42David C. LonieStatusbacklog => gatekeeper review
2014-02-11 12:42David C. LonieFixed in Version => git-next
2014-02-11 12:42David C. LonieResolutionopen => fixed
2014-02-11 13:40David C. LonieNote Added: 0032308
2014-02-11 13:40David C. LonieStatusgatekeeper review => active development
2014-02-19 09:05David C. LonieNote Added: 0032350
2014-02-19 09:05David C. LonieStatusactive development => gatekeeper review
2014-02-23 09:58Utkarsh AyachitFixed in Versiongit-next => git-master
2014-02-23 09:59Utkarsh AyachitStatusgatekeeper review => customer review
2014-02-23 09:59Utkarsh AyachitNote Added: 0032388
2014-02-24 13:03Utkarsh AyachitNote Added: 0032402
2014-02-24 13:03Utkarsh AyachitStatuscustomer review => todo
2014-03-03 14:44David C. LonieNote Added: 0032445
2014-03-03 15:11David C. LonieStatustodo => closed
2014-06-17 09:22Utkarsh AyachitTarget Version => 4.2
2014-08-29 11:00Utkarsh AyachitFixed in Versiongit-master => 4.2

Notes
(0032307)
David C. Lonie   
2014-02-11 12:42   
Added python objects to treat composite dataset arrays as a single array.
(0032308)
David C. Lonie   
2014-02-11 13:40   
Warnings/failures on dashboard
(0032350)
David C. Lonie   
2014-02-19 09:05   
Tests are passing on 'next' dashboards, should be good to go.
(0032388)
Utkarsh Ayachit   
2014-02-23 09:59   
SUMMARY
---------------------------------------------
Topics merged into master:
        14412-select-translucent-actors
        14455-save-screenshot-semicolon
        14539-can-crash
        14558-disconnect-cleanup
(VTK) 14565-composite-queries
        14585_XMLWriters_64bit_default
        ParallelPythonImport_mpi4py
        catalyst-web-improvement
        cleanup_CoProcessing
        fix-rendering-edition
        fix-sciberquest-installdir
        fix_multiview_widget
        pvweb-widget-catalyst
(0032402)
Utkarsh Ayachit   
2014-02-24 13:03   
Reopening due to a parallel bug. To reproduce do the following steps:

+ Run pvserver in parallel with 3 procs:
  mpirun -np 3 ./bin/pvserver
+ Connect to this pvserver using ParaView
+ Open can.ex2, check all variables, uncheck "Apply displacements", apply
+ Create D3 filter, apply. Color by "vtkProcessId" -- data partitioned among 3 procs.
+ Now open the Find Data dialog and try the query "EQPS[:,0] == max(EQPS)".

timestep 0: everything is fine (all cells have EQPS==0 so everything is highlighted).

timestep 1: incorrect -- if you look at the spreadsheet shown in the Find Data dialog, cells have EQPS = 0, 0.0013.., and 0.86.... That's not right.

(0032445)
David C. Lonie   
2014-03-03 14:44   
Applying the patch for 0013794 and using "global_max" instead of "max" works as expected.