Related to open issue #1595 and closed issue #1159
Downsampling in time fails when the downsampled UVData object has a scan_number_array parameter. The shape of the array is not updated after downsampling (Nblts axis keeps its size, ignoring the change in Ntimes), such that uvd.check() fails with ValueError: UVParameter _scan_number_array is not expected shape. I am trying to time-average a UVData object read from a measurement set, which has a _scan_number_array parameter, with uvd.downsample_in_time(n_times_to_avg=450). If I set uvd._scan_number_array to None, the error is not raised.
I think scan_number_array should be updated along with the other parameters of the UVData objects.
Thanks!