0

I am generating some netCDF datasets which contain daily variables over the 21st century, with each 5 years of data is stored in a separate file. Everything goes smoothly but when I checked the time values in each file I found out that for no obvious reason date2num generated some random big numbers (9.96921e+36) for some of the netCDF files! I am pretty sure that my dates array (which is converted to numbers) does not have any problem or missing data. In fact, I ran the same code again and it behaved for that specific file but generated random big numbers for another file. Has anyone come across this problem before?

Monobakht
  • 193
  • 1
  • 11
  • Can you provide a sample of the netcdf file along with the code you are using? That will help us diagnose the issue! – N1B4 Jun 09 '17 at 20:54
  • It turned out to be because of the shape of a variable that had time dimension. The variable X had dimensions of ('time','lat','lon') and its shape was (2501,200,200), while 'time' values were 2500 values. netCDF4 was not throwing any error about non-matching dimensions and was generating random numbers to match the size of time variable and first dimension of variable X. – Monobakht Jun 12 '17 at 10:31
  • The Rocket Software MultiValue Databases ( UniData, UniVerse and D3 ) use Python 3.4.1 not a 2.x version of Python, I think you may have put the wrong tag on the question. – Mike Jul 25 '17 at 23:43

0 Answers0