Error -90002 -

[Fotis Tsiolis]

Dear DREAM.3D team,

I am running into a persistent issue when trying to compute feature-averaged values of scalar fields that I appended to my DREAM.3D file.

Context:

• I start from a DAMASK simulation (.hdf5) and export orientations into a .dream3d file using damask library functionality, namely export_DREAM3D(q='O').

• After that, I append custom scalar arrays into /DataContainers/SyntheticVolumeDataContainer/CellData (namely: nuc_flag, tot_density, rho_mob_total, and rho_dip_total).

• Each array has the same TupleDimensions as Eulers, i.e. (478, 13, 108) with ComponentDimensions = [1].

• Example:

  • tot_density: shape (671112,), dtype float32

  • Metadata: TupleDimensions = [478, 13, 108], ComponentDimensions = [1]

When I open the file in DREAM.3D or in HDFView/myHDF5, the arrays are clearly present under CellData:

CellData: 

• EulerAngles,   
• Phases,   
• nuc_flag,   
• rho_dip_total,   
• rho_mob_total,   
• tot_density <-- array I want to average

I also confirm that FeatureIds exists after segmentation and that I select tot_density directly from the UI.

The problem:
When I add the Find Average Value of Scalars For Feature filter and point it to tot_density, the pipeline fails with:

The AttributeMatrix named 'CellData' does NOT have a DataArray with name 'tot_density'. This filter requires this DataArray in order to execute.

Pipeline output:

[6/10] Find Average Value of Scalars For Feature PIPLELINE FAILED

Yet, tot_density (and the other arrays) are still visible in the Data Structure pane after the Reader step.

Troubleshooting I tried:

• Verified in HDF5 browser → arrays are written correctly, with consistent tuple dimensions and types.

• Verified in DREAM.3D → arrays show up in the Data Structure pane after the Reader, and I selected them in the proxy.

• Confirmed that FeatureIds is created by segmentation.

• Tried disabling/re-enabling the filter, reordering the pipeline, and saving/reloading the file.

Observation:
The error only appears after segmentation. It looks as if the segmentation step may overwrite or recreate the CellData AttributeMatrix, causing DREAM.3D to “drop” arrays it doesn’t recognize.The appended scalar arrays should persist through segmentation and be available to subsequent filters (in particular, Find Average Value of Scalars For Feature).

Question:

• Is this a limitation/bug in DREAM.3D where segmentation replaces CellData and discards non-standard arrays?

• If so, is there a recommended workaround for keeping custom arrays (tot_density, etc.) alive through the segmentation pipeline?

Thank you very much in advance for your help.

Sincerely,

Fotis Tsiolis

--

Doctoral Candidate

Theory and Simulation Group

Department of Microstructure Physics and Alloy Design

Max-Planck-Institut für Nachhaltige Materialien

[Fotis Tsiolis]

You can fing my files in the following link, since i cannot upload them here.

CBS_3D_RVE-3_pass1_CBS_material_with_nuclea_flag__regridded_Cubic_0p8um_inc402.7z

[Michael Jackson]

I’m not sure what is wrong with that version of DREAM.3D, but I used DREAM3D-NX to process your pipeline and I was able to I think get what you wanted. Although I think we may have uncovered an issue in the 2D Charting module. 

I have attached the pipeline that was used in DREAM3D-NX

To directly address your concerns, NO filter in DREAM.3D version 6.5.171 should be deleting cell level arrays without warning you first. I would say there is a bug with that particular filter that was never resolved in that version. There is an alternate filter called “Compute Array Statistics” that I think is in that version of DREAM.3D if you want to stay with DREAM.3D. I think you would find DREAM3D-NX to be more feature rich. I leave that up to you.

Https://www.dream3d.io.

Academic licenses are free currently.

--
Mike Jackson                    mike.j...@bluequartz.net
BlueQuartz Software         www.bluequartz.net

President/Owner               Dayton, Ohio

Principal DREAM.3D Developer

--
You received this message because you are subscribed to the Google Groups "dream3d-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dream3d-user...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dream3d-users/b7785804-60bb-4112-b1fc-04e5b214a48cn%40googlegroups.com.

[Fotis Tsiolis]

Dear Mr. Jackson,

Thank you very much for your reply.

I managed to replicate the results you showed using Dream3D-NX, but the outputed data structure differs from the 6.x software version. 

As such, it is not readable from the recrystallization code I want to use, which was developed based on the 6.x data structure. 

I have also tried your suggestion regarding the "Find Attribute Array Statistics" Filter, but unfortunately I bumped into the same Error Code -90002 (see the attached screenshot).  

I have also tried the same pipelines with older 6.x binaries, but I am getting the same error. 

I guess there is something wrong with how I export my DAMASK hdf5 results into the D3D 6.x hdf5 structure that makes the segmentation pipeline fail. 

My question now is if there is a filter to convert from the -NX hdf5 data structure back to the 6.x hdf5 structure.

Thank you very much in advance for your help.

Sincerely,

Fotis Tsiolis

--

Doctoral Candidate

Theory and Simulation Group

Department of Microstructure Physics and Alloy Design

Max-Planck-Institut für Nachhaltige Materialien

[Michael Jackson]

Dear Fotis,

    DREAM3D-NX does not have a filter to write out a DREAM.3D 6.x compatible file. Whose recrystallization codes do you use? Is it possible to update those codes to read the new HDF5 layout? Or is there an alternate input that the recrystallization codes can read that DREAM3D-NX could write out?

--
Mike Jackson                    mike.j...@bluequartz.net
BlueQuartz Software         www.bluequartz.net

President/Owner               Dayton, Ohio

Principal DREAM.3D Developer

On Sep 18, 2025 at 10:07:27, Fotis Tsiolis <fotis...@hotmail.gr> wrote:

Dear Mr. Jackson,

Thank you very much for your reply.

I would like to infrom you that I was able to replicate your segmentation pipeline results using the new -NX version of D3D. I have noticed thought that the resultant hdf5 structure differs from the 6.x version of the software. This causes segmentation errors with the recrystallization code in which I want to use the segmented RVE ans an input. Therefore, I also tried the "Finf Attribute Arrays Statistics" filter that you mentioned, but I bumped into the same error code -90002 (see the screenshot attached).

Moreover, I used an older 6.x version than the last one that I currrently use to test my segmentation pipeline there. Unfortunately, the error persists. I guess there is something wrong with the python code I use to export my results from the DAMASK hdf5 file to a 6.x  compatible .dream3d hdf5 file.

Therefore, I would like to ask if there is a filter that could write the segmentation results into the 6.x hdf5 data fromat.

Thank you very much in advance for your help.

Sincerely,

Fotis Tsiolis

--

Doctoral Candidate

Theory and Simulation Group

Department of Microstructure Physics and Alloy Design

Max-Planck-Institut für Nachhaltige Materialien

To view this discussion visit https://groups.google.com/d/msgid/dream3d-users/b939ce6c-c750-41e1-b001-8871ab4e25b1n%40googlegroups.com.

[Michael Jackson]

Dear Fotis, 

    Without having the original DAMASK data it is hard to say what is going wrong. I would like to point out that both DREAM.3D and DREAM3D-NX have a filter to directly import data from an HDF5 file. You will have to do add some filters to setup the proper geometry, but then you can just directly read the data from the DAMASK HDF5 file into DREAM3D-NX (or DREAM.3D if you really want to stay with that version).

I would also like to point out https://www.dream3d.io/python_docs which is for DREAM3D-NX but gives you access to all the filters in DREAM3D-NX through python scripting. You can also create DREAM3D-NX filters written in Python so one could write a DAMASK file importer using python.

If you have an example DAMASK output file that represents everything that DAMASK outputs I would be happy to take a look and offer some helpful ideas on how to easily get the data into DREAM3D-NX for processing.

--
Mike Jackson                    mike.j...@bluequartz.net
BlueQuartz Software         www.bluequartz.net

President/Owner               Dayton, Ohio

Principal DREAM.3D Developer

On Sep 18, 2025 at 10:18:27, Fotis Tsiolis <fotis...@hotmail.gr> wrote:

Dear Mr. Jackson,

Thank you for your detailed reply and for testing the pipeline in DREAM3D-NX. I was able to replicate the results with NX as well. However, I noticed that the data structure in NX differs from the 6.x versions, which are the ones compatible with the recrystallization code I am currently using.

Following your suggestion, I also tested the Compute Array Statistics filter in 6.x, but unfortunately, I still received the same -90002 error. I even tried running the workflow with previous 6.x versions, but the issue persisted. At this point, I suspect the problem may be related to the Python script I wrote to export DAMASK HDF5 data into a 6.x-compatible HDF5 format for segmentation of my RVE.

Before I start developing a more robust conversion script, I would like to ask if DREAM.3D provides any filter or functionality to write datasets back into the legacy HDF5 format (6.x compatible). This would save some time compared to building a custom exporter.

Thanks again for your support.

Best regards,

Fotis Tsiolis

--

Doctoral Candidate

Theory and Simulation Group

Department of Microstructure Physics and Alloy Design

Max-Planck-Institut für Nachhaltige Materialien

Στις Πέμπτη 18 Σεπτεμβρίου 2025 στις 1:25:01 π.μ. UTC+2, ο χρήστης Michael Jackson έγραψε:

[Fotios Tsiolis]

Dear Mr. Jackson,

Thanks for the follow-up and for pointing me to the HDF5 import functionality. I am aware of the HDF5 reader in DREAM.3D 6.x and will also give it another try, but in principle, it should behave similarly to the Python script I wrote, which outputs a DREAM.3D-compatible HDF5 file.

For your reference, here is an example DAMASK HDF5 file that reflects the data I am working with:

​CBS_3D_RVE-3_pass1_CBS_material_with_nuclea_flag.hdf5​

In parallel, I will be in contact with the recrystallization code developers to evaluate which route is more suitable moving forward:

1. Adapting the RX code to work with DREAM3D-NX outputs, or
2. Writing a script to revert DREAM3D-NX outputs back to the legacy 6.x-compatible format.

I will keep you posted as this develops.

Thanks again for your support and for offering to take a look at the DAMASK file.

Best regards,
Fotis

---

Από: Michael Jackson <mike.j...@bluequartz.net>
Στάλθηκε: Παρασκευή, 19 Σεπτεμβρίου 2025 3:46 μμ
Προς: Fotis Tsiolis <fotis...@hotmail.gr>
Κοιν.: dream3d-users <dream3...@googlegroups.com>
Θέμα: Re: Προσωπικό μήνυμα σχετικά με: Error -90002 -

 

[Michael Jackson]

Dear Fotios, 

This is a start on the documentation for the latest DREAM3D-NX file specification.

    https://www.dream3d.io/nx_reference_manual/html/Concepts/DREAM3DFileSpecification.html. This might help the DAMASK folks write a file.

I will try to take a look at the supplied file soonish but due to my schedule it might be a few weeks.

--
Mike Jackson                    mike.j...@bluequartz.net
BlueQuartz Software         www.bluequartz.net

President/Owner               Dayton, Ohio

Principal DREAM.3D Developer

[Michael Jackson]

Attached is the pipeline that I used to import the data. Due to the use of Compound datasets in the HDF5 file there are a few datasets that cannot be imported. We are looking into supporting these types in a future version of DREAM3D-NX. 

--
Mike Jackson                    mike.j...@bluequartz.net
BlueQuartz Software         www.bluequartz.net

President/Owner               Dayton, Ohio

Principal DREAM.3D Developer

[Fotios Tsiolis]

Dear Mr. Jackson,

Thank you very much for having a look at my data. 

It seems that the HDF5 importing of the DAMASK files is similar with how such datasets are read in the 6.x version. 

Maybe a DAMASK reader from your side or updating of DAMASK's D3D exporter could be something worth exploring in the -NX version of the software.

Best regards, 

Fotis Tsiolis

--

Doctoral Candidate

Theory and Simulation Group

Department of Microstructure Physics and Alloy Design

Max-Planck-Institut für Nachhaltige Materialien

Max-Planck-Str. 1, D-40237, Düsseldorf

Office: H2 5L1, E-mail: f.ts...@mpi-susmat.de

---

Από: Michael Jackson <mike.j...@bluequartz.net>
Στάλθηκε: Δευτέρα, 22 Σεπτεμβρίου 2025 7:24 μμ
Προς: Fotios Tsiolis <fotis...@hotmail.gr>