One of the most useful features of hdf5 is it's ability to hold big/multidimensional data - this allows you to store complicated multi-modal scans (like combined transmission, xrf and xrd tomography) in a way that reflects the measurement better than a column.
February 17, 2025 at 1:05 PM
One of the most useful features of hdf5 is it's ability to hold big/multidimensional data - this allows you to store complicated multi-modal scans (like combined transmission, xrf and xrd tomography) in a way that reflects the measurement better than a column.
Yeah, we have discussed this for many hours at Diamond. There is definitely a set of measurements for which an NXdata can map quite nicely to a pandas dataframe, but when you try to generalise over all the experiments done at a synchrotron it's not that large a set.
February 17, 2025 at 12:02 PM
Yeah, we have discussed this for many hours at Diamond. There is definitely a set of measurements for which an NXdata can map quite nicely to a pandas dataframe, but when you try to generalise over all the experiments done at a synchrotron it's not that large a set.
Ah yes, if it can read the data, but it is not in the correct structure that is a different problem. If you wanted to go direct from Nexus to Data frame you would probably need a specific Nexus base class/application definition (or custom import code).
February 17, 2025 at 10:54 AM
Ah yes, if it can read the data, but it is not in the correct structure that is a different problem. If you wanted to go direct from Nexus to Data frame you would probably need a specific Nexus base class/application definition (or custom import code).