# Using Point Data as Cell Centers in the Surface With Edges Visualisation

**URL:** https://discourse.paraview.org/t/using-point-data-as-cell-centers-in-the-surface-with-edges-visualisation/12350
**Category:** ParaView Support
**Tags:** vtk
**Created:** [June 20, 2023, 2:47pm UTC](https://discourse.paraview.org/t/using-point-data-as-cell-centers-in-the-surface-with-edges-visualisation/12350 "2023-06-20T14:47:41Z")
**Posts on this page:** 4
**Page:** 1

<div class="post-metadata">

### Author: ![grld](https://discourse.paraview.org/letter_avatar_proxy/v4/letter/g/c77e96/32.png) [@grld](https://discourse.paraview.org/u/grld)
#### Post date: [June 20, 2023, 2:47pm UTC](https://discourse.paraview.org/t/using-point-data-as-cell-centers-in-the-surface-with-edges-visualisation/12350/1 "2023-06-20T14:47:41Z")

</div>

Dear community,

I would like to use ParaView to inspect the results I obtain from a cell-centered solver. The results obtained from the solver correspond to the midpoints of cubic volumes. However, ParaView interprets them as the vertex values of a structured grid. When choosing “Surface With Edges” as visualization this becomes obvious. In that case, I obtain a volume with 21 cubic cells in each direction instead of 22. I have visualized my problem in the enclosed figure. The dark points are the raw data that are interpreted as vertice values of a structured grid. The pink values would be the vertices of the larger grid, in which the raw data corresponds to the cell centers.

I am looking for a way to tell ParaView that the source data are the cell midpoints instead of the cell vertices. Is there any straightforward way to do this? I guess it requires extrapolation from the cell centers to the vertices and an averaging step at those.

Thank you very much!

 ![ParaView_Grid](https://discourse.paraview.org/uploads/default/original/2X/b/b97e7bbc35074b6cbac44d8076e82e5826d91ca9.png)

---

<div class="post-metadata">

### Author: ![Kenichiro-Yoshimi](https://discourse.paraview.org/letter_avatar_proxy/v4/letter/k/ecc23a/32.png) [@Kenichiro-Yoshimi](https://discourse.paraview.org/u/Kenichiro-Yoshimi)
#### Post date: [June 20, 2023, 11:35pm UTC](https://discourse.paraview.org/t/using-point-data-as-cell-centers-in-the-surface-with-edges-visualisation/12350/2 "2023-06-20T23:35:21Z")

</div>

For image data (Uniform Rectilinear Grid), it may be a good idea to save the data once to Xdmf Data File (File \> Save Files… \> Xdmf Data File (\*.xmf)) and modify the origin position, Dimensions of Topology, and Center of Attribute in the xmf file as follows.

[sample.zip](https://discourse.paraview.org/uploads/short-url/4MPOqjYfcGQTVqK8GteTDSHUXtZ.zip) (6.7 KB)

- org.xmf

```xml
<?xml version="1.0" encoding="utf-8"?>
<Xdmf xmlns:xi="http://www.w3.org/2001/XInclude" Version="3.0">
  <Domain>
    <Grid Name="Grid">
      <Geometry Origin="" Type="ORIGIN_DXDYDZ">
        <DataItem DataType="Float" Dimensions="3" Format="XML" Precision="8">-10 -10 -10</DataItem>
        <DataItem DataType="Float" Dimensions="3" Format="XML" Precision="8">1 1 1</DataItem>
      </Geometry>
      <Topology Dimensions="21 21 21" Type="3DCoRectMesh"/>
      <Attribute Center="Node" ElementCell="" ElementDegree="0" ElementFamily="" ItemType="" Name="DistanceSquared" Type="None">
        <DataItem DataType="Float" Dimensions="21 21 21" Format="HDF" Precision="4">org.h5:Data0</DataItem>
      </Attribute>
    </Grid>
  </Domain>
</Xdmf>

```

- cell\_centerd.xmf

```xml
<?xml version="1.0" encoding="utf-8"?>
<Xdmf xmlns:xi="http://www.w3.org/2001/XInclude" Version="3.0">
  <Domain>
    <Grid Name="Grid">
      <Geometry Origin="" Type="ORIGIN_DXDYDZ">
        <DataItem DataType="Float" Dimensions="3" Format="XML" Precision="8">-10.5 -10.5 -10.5</DataItem>
        <DataItem DataType="Float" Dimensions="3" Format="XML" Precision="8">1 1 1</DataItem>
      </Geometry>
      <Topology Dimensions="22 22 22" Type="3DCoRectMesh"/>
      <Attribute Center="Cell" ElementCell="" ElementDegree="0" ElementFamily="" ItemType="" Name="DistanceSquared" Type="None">
        <DataItem DataType="Float" Dimensions="21 21 21" Format="HDF" Precision="4">org.h5:Data0</DataItem>
      </Attribute>
    </Grid>
  </Domain>
</Xdmf>

```

This modification would change the data as follows:

![sample](https://discourse.paraview.org/uploads/default/original/2X/2/2c5e99291f8692f27892f012b70d6b432dd3dee1.gif)

---

<div class="post-metadata">

### Author: ![grld](https://discourse.paraview.org/letter_avatar_proxy/v4/letter/g/c77e96/32.png) [@grld](https://discourse.paraview.org/u/grld)
#### Post date: [June 21, 2023, 1:42pm UTC](https://discourse.paraview.org/t/using-point-data-as-cell-centers-in-the-surface-with-edges-visualisation/12350/3 "2023-06-21T13:42:01Z")

</div>

Thank you @Kenichiro-Yoshimi for the fast and helpful reply!  
I tried to follow your instructions but in my case, the .xmf file looks a little different in terms of Geometry, Topology. It already had the “Dimensions=22 22 22” in the Topology.

```auto
<?xml version="1.0" encoding="utf-8"?>
<Xdmf xmlns:xi="http://www.w3.org/2001/XInclude" Version="3.0">
  <Domain>
    <Grid Name="Grid">
      <Geometry Origin="" Type="XYZ">
        <DataItem DataType="Float" Dimensions="10648 3" Format="HDF" Precision="8">PhaseField_0.h5:Data2</DataItem>
      </Geometry>
      <Topology Dimensions="22 22 22" Type="3DSMesh"/>
      <Attribute Center="Node" ElementCell="" ElementDegree="0" ElementFamily="" ItemType="" Name="PhaseFields" Type="None">
        <DataItem DataType="Int" Dimensions="22 22 22" Format="HDF" Precision="4">PhaseField_0.h5:Data3</DataItem>
      </Attribute>
    </Grid>
  </Domain>
</Xdmf>

```

Should I replace the Geometry type and the Topology type of my data to match your version?

---

<div class="post-metadata">

### Author: ![Kenichiro-Yoshimi](https://discourse.paraview.org/letter_avatar_proxy/v4/letter/k/ecc23a/32.png) [@Kenichiro-Yoshimi](https://discourse.paraview.org/u/Kenichiro-Yoshimi)
#### Post date: [June 22, 2023, 12:19am UTC](https://discourse.paraview.org/t/using-point-data-as-cell-centers-in-the-surface-with-edges-visualisation/12350/4 "2023-06-22T00:19:37Z")

</div>

Since your data type is not Image but `Structured (Curvilinear) Grid`, you cannot directly apply the method I suggested. If the spacing of the grid is constant, you could just change it to `Image (Uniform Rectilinear Grid)` data type by using the `Resample To Image` filter and then modify the xmf file, but that seems tedious. I am not sure what format the solver currently uses, but I think it would be better to output the data directly as cell center data.