# Connectivity filter for parallel files (PVTU)

**URL:** https://discourse.paraview.org/t/connectivity-filter-for-parallel-files-pvtu/5595
**Category:** ParaView Support
**Created:** [October 19, 2020, 8:32am UTC](https://discourse.paraview.org/t/connectivity-filter-for-parallel-files-pvtu/5595 "2020-10-19T08:32:37Z")
**Posts on this page:** 5
**Page:** 1

<div class="post-metadata">

### Author: ![JayghoshRao](https://discourse.paraview.org/user_avatar/discourse.paraview.org/jayghoshrao/32/5151_2.png) [@JayghoshRao](https://discourse.paraview.org/u/JayghoshRao)
#### Post date: [October 19, 2020, 8:32am UTC](https://discourse.paraview.org/t/connectivity-filter-for-parallel-files-pvtu/5595/1 "2020-10-19T08:32:37Z")

</div>

Hi,

My dataset consists of multiple spheres within a large cylinder. I need to integrate scalars within the individual spheres and extract results. I can do this in serial using the following workflow:

1. Open
2. Connectivity Filter
3. Threshold Filter
4. Integrate Variables Filter

But this gives me bad results when dealing with parallel vtk files (PVTU), thanks to the split vtu files counting as separate regions.

Currently, there are no ghost elements in the VTU files that I use.

Is there a way to get the connectivity filter to ignore the vtu splits? I tried using cleanToGrid but it ends up merging everything into a monolithic cylinder. Is there a way to merge ONLY the partitioned vtu data from pvtu files?

Thanks!

---

<div class="post-metadata">

### Author: ![cory.quammen](https://discourse.paraview.org/user_avatar/discourse.paraview.org/cory.quammen/32/11193_2.png) [@cory.quammen](https://discourse.paraview.org/u/cory.quammen)
#### Post date: [October 19, 2020, 1:34pm UTC](https://discourse.paraview.org/t/connectivity-filter-for-parallel-files-pvtu/5595/2 "2020-10-19T13:34:10Z")

</div>

ParaView 5.9 will add a mode to the **Merge Blocks** filter that merges only pieces in a multipiece data set. That should do what you are looking for if I am not mistaken. @utkarsh.ayachit, can you verify that will apply in this case?

---

<div class="post-metadata">

### Author: ![JayghoshRao](https://discourse.paraview.org/user_avatar/discourse.paraview.org/jayghoshrao/32/5151_2.png) [@JayghoshRao](https://discourse.paraview.org/u/JayghoshRao)
#### Post date: [October 19, 2020, 2:24pm UTC](https://discourse.paraview.org/t/connectivity-filter-for-parallel-files-pvtu/5595/3 "2020-10-19T14:24:08Z")

</div>

Sounds about right, but I’m clueless about the implementation. But Merge Blocks doesn’t apply to pvtu files, right?

Does this mean the procedure is:

1. Open PVTU in Parallel
2. ProcessIds
3. Threshold for each ID
4. Merge Blocks

or is there a way to make it directly apply to a PVTU file?

Also, is there a date for the 5.9 release?

Thanks, again!

---

<div class="post-metadata">

### Author: ![cory.quammen](https://discourse.paraview.org/user_avatar/discourse.paraview.org/cory.quammen/32/11193_2.png) [@cory.quammen](https://discourse.paraview.org/u/cory.quammen)
#### Post date: [October 19, 2020, 2:32pm UTC](https://discourse.paraview.org/t/connectivity-filter-for-parallel-files-pvtu/5595/4 "2020-10-19T14:32:34Z")

</div>

D’oh, right. Sorry, Merge Blocks only applies to multiblock datasets. I suppose you could use the **Group** filter, then **Merge Blocks** with the single-block dataset, then **Extract Block** to get it out. That’s a pretty ugly workflow, though. A separte **Merge Pieces** filter would be easier to use.

Date of the 5.9 release is slated for sometime before Supercomputing 2020 (there is no set date).

---

<div class="post-metadata">

### Author: ![JayghoshRao](https://discourse.paraview.org/user_avatar/discourse.paraview.org/jayghoshrao/32/5151_2.png) [@JayghoshRao](https://discourse.paraview.org/u/JayghoshRao)
#### Post date: [January 28, 2021, 6:09am UTC](https://discourse.paraview.org/t/connectivity-filter-for-parallel-files-pvtu/5595/5 "2021-01-28T06:09:04Z")

</div>

Hi Cory,

Just following up on any possible updates. Has something like **Merge Pieces** been implemented yet?

To reiterate, I have a cylinder filled with spheres, and I’d like to use **Integrate Variables** on every sphere. Extracting the sphere is possible in serial mode using the **Connectivity** filter. However, if the data is distributed over several files, the **Connectivity** filter separates out the partitions, and not the spheres from the cylinders.

Additionally, is it possible to apply **Integrate Variables** in parallel over multiple objects? Currently I resort to using **Threshold** to individually separate and loop over every object.
