Hotfix data_set.py: flatten 2D arrays

[GOTYGuyOfTheYear]

When adding a 2D array as a standalone parameter in a dataset, the whole numpy array would be save in bit form. Flatten is needed to make a proper array to be saved alongside the data.

[jenshnielsen]

When adding a 2D array as a standalone parameter in a dataset, the whole numpy array would be save in bit form. Flatten is needed to make a proper array to be saved alongside the data.

I am not sure that I follow what the issue is. When a user stores an array of a given shape I would imagine that they would also expect this to be read back.

Could you perhaps supply an example or a test that shows the issue that you are facing?

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 50.00000% with 1 line in your changes missing coverage. Please review.

Project coverage is 69.34%. Comparing base (92e4d95) to head (8f4baa8). Report is 27 commits behind head on main.

Files with missing lines	Patch %	Lines
src/qcodes/dataset/data_set.py	50.00%	1 Missing :warning:

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #6573   +/-   ##
=======================================
  Coverage   69.34%   69.34%           
=======================================
  Files         340      340           
  Lines       31266    31266           
=======================================
  Hits        21681    21681           
  Misses       9585     9585           

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[GOTYGuyOfTheYear]

Hello @jenshnielsen,

Thank you for your interest. The issue appears when value is a 2D numpy array. I guess the python line : [{param.name: number} for number in value] interates through the columns of the array. When we try to load the array and convert it to xarray again we get this error:

In the sqlite db, the fields for the independent parameter are filled with a blob type:

There might be a better fix or a better way of working, this is just how I have found as the simplest option.