ValueError is raised, if cycle_nr in channel group is lower than actual data-length

[janriem]

Hi Daniel,

starting with version 8.1.0, there is an exception, when reading from a channel group whose cycle_nr is wrong (too low) in the file.
Before this version the behaviour was to read cycle_nr samples from the signal (i.e. cutting the signal short) without failing.

The exception message "ValueError: memoryview assignment: lvalue and rvalue have different structures" is raised by this line:

asammdf/src/asammdf/blocks/mdf_v4.py

Line 1784 in 9a9f719

buffer_view[cur_size : cur_size + original_size] = new_data

Although this is caused by an error in the file, I wanted to ask, if you could restore the previous behaviour so reading the signal doesn't fail?
If there was a way to read the full signal, that would of course be even better.

Example Code:

mdf = MDF("cycle_count.mf4")
signal = mdf.get("RandomSignal")

Versions:

• tested with latest development version (https://github.com/danielhrisca/asammdf/tree/9a9f719160a199842e75453a55a71aa9f28d2bdc)
• behaviour changed with this commit: 71f03b7

Testfile:

cycle_count.mf4.zip

---

Jan Riemer jan.riemer@mercedes-benz.com on behalf of Mercedes-Benz Tech Innovation GmbH
https://github.com/mercedes-benz/foss/blob/master/PROVIDER_INFORMATION.md

[danielhrisca]

Hello Jan,

this branch code https://github.com/danielhrisca/asammdf/tree/scale_axis seems to work correctly with your file

[janriem]

Hello Daniel,

I looked into it and with this branch I do get the shorter (cut-off) signal without an exception.
What are your plans regarding the scale_axis branch? Is it going to be merged into development and released?

[danielhrisca]

Hello Jan, I've merged scale_axis into development. I will need to make more documentation updates before making a new release