[Bug]: sds-convert does not properly handle different value types

[MiloradCvjetkovic]

Is there an existing issue for this?

• I have searched the existing issues

Problem statement

If an SDS record contains different value types then sds-convert fails with message like below:

sds-convert simple_csv -i ML_Out.0.sds -o ML_Out.0.csv -y ML_Out.sds.yml
Traceback (most recent call last):
  File "D:\Repositories\GitHub\Arm-Examples\SDS-Examples\Alif\AppKit-E7_USB\algorithm\SDS Recordings\sds-convert.py", line 837, in <module>
    main()
    ~~~~^^
  File "D:\Repositories\GitHub\Arm-Examples\SDS-Examples\Alif\AppKit-E7_USB\algorithm\SDS Recordings\sds-convert.py", line 799, in main
    write_SDS_SimpleCSV(args, data[sensor_name[0]], meta_data[sensor_name[0]])
    ~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\Repositories\GitHub\Arm-Examples\SDS-Examples\Alif\AppKit-E7_USB\algorithm\SDS Recordings\sds-convert.py", line 287, in write_SDS_SimpleCSV
    tmp_record_row = [record_row[i][sample_idx] for i in range(len(record_row))]
                      ~~~~~~~~~~~~~^^^^^^^^^^^^
IndexError: list index out of range

Steps To Reproduce

Use the files in the repository and run:

sds-convert simple_csv -i ML_Out.0.sds -o ML_Out.0.csv -y ML_Out.sds.yml

[VladimirUmek]

Following instructions I was not able to reproduce the issue using latest convert tool (from this repository). Can you check again please?

[MiloradCvjetkovic]

Indeed the error is not present anymore, however there is another problem with the conversion which can still be considered under same topic, and it is:

The conversion is not correct.

The metadata ML_Out.sds.yml contains double, uint32_t, uint32_t, uint32_t, uint32_t but the result in the csv file seems to not reflect that correctly.