Dear all,
is there support for different types of NaNs? I need to be able to store
dataset of real numbers. The data itself may contain NaN values, it is OK.
But there can be cells without data (which has not been calculated so far)
so I need to put some distinguishable value, but NaN is already used. We
are thinking about using 'NaN' with payload for that purpose, but we did
not try yet. So my question is if there is support for such types of NaN in
HDF5?
Best regards,
Alexander.
11.09.2014 18:18, Alexander Prokhorov пишет:
Dear all,
is there support for different types of NaNs? I need to be able to store
dataset of real numbers. The data itself may contain NaN values, it is
OK. But there can be cells without data (which has not been calculated
so far) so I need to put some distinguishable value, but NaN is already
used. We are thinking about using 'NaN' with payload for that purpose,
but we did not try yet. So my question is if there is support for such
types of NaN in HDF5?
Hello Alexander!
I believe Fill Values (http://www.hdfgroup.org/HDF5/doc_resource/H5Fill_Values.html) is the way to encode missing data. As for present but invalid data, it seems that you can put NaNs into the dataset without problems.
Best wishes,
Andrey Paramonov
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
