# Discontinuity in output of GradientOfUnstructuredDataSet
**URL:** https://discourse.paraview.org/t/discontinuity-in-output-of-gradientofunstructureddataset/1738
**Category:** ParaView Support
**Created:** [April 29, 2019, 6:47pm UTC](https://discourse.paraview.org/t/discontinuity-in-output-of-gradientofunstructureddataset/1738 "2019-04-29T18:47:04Z")
**Posts on this page:** 2
**Page:** 1
### Author:  [@brenerrr](https://discourse.paraview.org/u/brenerrr)
#### Post date: [April 29, 2019, 6:47pm UTC](https://discourse.paraview.org/t/discontinuity-in-output-of-gradientofunstructureddataset/1738/1 "2019-04-29T18:47:04Z")
I currently have a snapshot of a CFD simulation in CGNS format with 10 zones as shown below.
When I calculate Q-Criterion with the GradientOfUnstructuredDataSet filter, discontinuities appear at the zones’ boundaries. I’m pretty sure that’s happening because paraview is not understanding that it should use points from the next (or previous) zone in the derivative calculations, even though the CGNS file does have a GridConnectivity node, and ends up treating that location as a boundary.
Does anyone know how I could inform paraview to use proper points from other blocks of data or something in this sense that could fix my issue?
Thank you in advance
---
### Author:  [@brenerrr](https://discourse.paraview.org/u/brenerrr)
#### Post date: [May 2, 2019, 3:33pm UTC](https://discourse.paraview.org/t/discontinuity-in-output-of-gradientofunstructureddataset/1738/2 "2019-05-02T15:33:32Z")
Well, in case someone stumbles into this problem, I solved it by creating a new vtkStructuredGrid with all the blocks appended. I was able to do that with a programmable filter.