With vlen datatypes, writing to the dataset would raise an error if the vlen types were uniform. i.e., an array like:
[ndarray[1, 2], ndarray[3,4], ndarray[5,6]] would be processed as [[1,2],[3,4],[5,6]] (a shape of (3,2) during the arrayToBytes conversion rather than (3), causing a broadcast error. This PR fixes that so the data can be written to the expected shape.
With vlen datatypes, writing to the dataset would raise an error if the vlen types were uniform. i.e., an array like:
[ndarray[1, 2], ndarray[3,4], ndarray[5,6]] would be processed as [[1,2],[3,4],[5,6]] (a shape of (3,2) during the arrayToBytes conversion rather than (3), causing a broadcast error. This PR fixes that so the data can be written to the expected shape.