Hi Billy,
How about using the embedded _expression_ `var,return=tunits`? You'd have to do your own conversion from days->seconds or whatever via an IF-ELSE block.
Cheers,
Russ
From: owner-ferret_users@xxxxxxxx <owner-ferret_users@xxxxxxxx> on behalf of William Kessler <william.s.kessler@xxxxxxxx>
Sent: Monday, 2 December 2019 9:20 AM To: Ferret <ferret_users@xxxxxxxx> Cc: William S. Kessler <william.s.kessler@xxxxxxxx> Subject: [ferret_users] How to find the number of seconds in a timestep I'm sure I'm just missing something here ....
I'm reading variables from a netcdf file that has a regular time axis in days (as I know by listing a few timesteps). I can also see by this listing that the timesteps are 2 hours, but neither the axis unit nor the timestep length is a number I can use (without listing and looking each time). How can I reliably find the number of seconds in a timestep? Alternately, how can I find the units of the time axis? yes? sho axis time name axis # pts start end TIME TIME 7149 r 17-JUL-2012 04:00 04-MAR-2014 20:00 T0 = 01-JAN-1970 00:00:00 let TT = t[gt=u] ! listing a few TT values shows by eye that the time axis is days, and the timestep 2 hours, but these are not numbers I can use except manually list TBOX[gt=u] ! shows that TBOX=1/12 of the time axis unit, but I don't know that unit length except by eye-balling a listing let TAXIS_UNIT = 1/TT[l=@ddf] ! length of a time axis unit in seconds (because Ferret takes time derivatives in seconds) let TSEC = TBOX[gt=u] * TAXIS_UNIT ! this calculation finally tells me the length of a timestep, in seconds Clearly Ferret knows this information but won't tell me without a calculation. Am I missing something obvious? It seems that TAXIS_UNIT should be a keyword like TBOX, TBOXLO, TBOXHI. Those are all similar to index units in that they are only meaningful once you know what the geophysical time axis unit is. TAXIS_UNIT (or equivalently TSEC) would make these available for calculations. And note that my method above fails at the final timestep (because of @ddf). In the present case that doesn't matter because I know that the axis is regular, but if it wasn't I would have a hard time getting this information. Thanks ... Billy K |