# How to filter small droplets
**URL:** https://discourse.paraview.org/t/how-to-filter-small-droplets/486
**Category:** ParaView Support
**Created:** [August 28, 2018, 6:16pm UTC](https://discourse.paraview.org/t/how-to-filter-small-droplets/486 "2018-08-28T18:16:17Z")
**Posts on this page:** 5
**Page:** 1
### Author:  [@DavidSCN](https://discourse.paraview.org/u/DavidSCN)
#### Post date: [August 28, 2018, 6:16pm UTC](https://discourse.paraview.org/t/how-to-filter-small-droplets/486/1 "2018-08-28T18:16:17Z")
Hello everyone,
we run simulations of secondary atomization of water drops.
In this simulations, parts of the drop separate from the “main” drop.
I need to filter out these small drop separation, in order to get the quantities of interest, because all quantities refer to the main drop.
I have also attached a figure, to illustrate the problem.
It is easy to filter the drops, itself, because they have a much higher density (ambient flow is air).
But I could not find any solution to filter the main and the small drops.
I calculated the volume and tried to use this as a criterion, but without any success. I think the solution could be something like the iteration from an initial cell on the center line, but here, I could also not implement something similar.
Maybe someone here has any suggestion, how to handle the problem. Of course you can do it manually by selecting the cells, but there are hundreds of data files and this is simply not possible,
Any help and or idea is appreciated.
Best regards,
David
---
### Author:  [@nicolas.vuaille](https://discourse.paraview.org/u/nicolas.vuaille)
#### Post date: [August 29, 2018, 8:36am UTC](https://discourse.paraview.org/t/how-to-filter-small-droplets/486/2 "2018-08-29T08:36:16Z")
Hi,
You can try in two steps: extract all drops with a `Clip` filter (set type to `scalar` and use the denisty) and then extract the main drop using the `Connectivity` filter, with the mode `Extract largest region`.
Regards,
---
### Author:  [@DavidSCN](https://discourse.paraview.org/u/DavidSCN)
#### Post date: [August 29, 2018, 1:56pm UTC](https://discourse.paraview.org/t/how-to-filter-small-droplets/486/3 "2018-08-29T13:56:51Z")
Hi Nicolas,
many thanks for your suggestion. In principle, this is exactly what I need. The problem: We use a Multiresolution in the algorithm (similar to an AMR), but the AMR Filters are not accessible. It seems like the `Connectivity` Filter is based on the region\_id. By applying this filter, I only get a small block of the (multiresoluted) block.
In the figure, the tiny block at the bottom left in red is the ‘largest region’. The connectivity filter dosn’t allows me to choose a criterion for the connectivity. I think therefore I cannot use it proper-
Best regards.
PS: The image refers to the legend of the region\_ids, I forgot to remove the second view of the largest region.
EDIT:
The discontinuities in the region id are a result of the parallel computation of the simulation. It corresponds to the partition.
---
### Author:  [@nicolas.vuaille](https://discourse.paraview.org/u/nicolas.vuaille)
#### Post date: [August 29, 2018, 2:09pm UTC](https://discourse.paraview.org/t/how-to-filter-small-droplets/486/4 "2018-08-29T14:09:45Z")
Hi,
I’m not very familiar with multiresolution … Can you use a `Merge Blocks` filter before the `Connectivity` ?
If you have some skills in python, you may also try the `Programmable` filter. See [https://www.paraview.org/Wiki/Python\_Programmable\_Filter](https://www.paraview.org/Wiki/Python_Programmable_Filter)
---
### Author:  [@DavidSCN](https://discourse.paraview.org/u/DavidSCN)
#### Post date: [August 29, 2018, 4:52pm UTC](https://discourse.paraview.org/t/how-to-filter-small-droplets/486/5 "2018-08-29T16:52:03Z")
Unfortunately not. Probably you are right, in need to get in touch with the `Programmable Filter`. But ven there, it seems that I cannot steer the way paraView iterates over the different cells to create a kind of connectivity/continuity criteria. But thank you very much anyway.
Best regards!