BenK

Hello besides I can hardly discuss on a deeper developer level like it was in the middle of this discussion, I'd like to give some feedback to the initial question: Yes, me. No, seriously: In my opinion, netCDF4 has - despite "only" being some kind of fork on top of HDF5 - some conceptually important differences, which are for me personally the reason to choose it again and again before HDF5. (I looked some times towards HDF5 because of the existing LV-interface, but I always found it to be too general for my needs.) The data I receive from my measurement device is n-dimensional, while n can differ from measurement to measurement. This was the main reason I was searching for something different than tdms or just ASCII-files. Important for my point of view is that the independent variables of such a dataset are coordinates and I want that information to be stored in the data file. Perhaps this is best described here http://www.unidata.ucar.edu/blogs/developer/entry/dimensions_scales. So what I did was running the LV-assistant for creating wrapper-VIs for given c-headers and the adjusting manually step by step the things to be made that it runs without errors... But I have to admit that this usage of netCDF4 has not made it into my main every days LV-project which runs my measurement device for years now. It's still part of the new beta-like project - but it works and produces fine nc-files.