Re: [ferret_users] accuracy with axis in seconds

[Brockmann Patrick <Patrick.Brockmann@xxxxxx>]

Ansley Manke a écrit :
> Hi Patrick,
> We had another idea for setting up your log.  See if this works out 
> for you.  Define an axis ending on the exact date and time you want to 
> measure to, and use `tt,return=Lsize` to get the length of the axis. 
> Then redefine the axis with another end date and get that length.
>
>     define symbol time_end = 02-JUN-2010 13:34:45
>
>     define
>     axis/t="1-jan-2008":"($time_end)":1/units=seconds/t0="1-jan-2008"/cal=gregorian
>     mytaxis
>      !-> def axis/t="1-jan-2008":"02-JUN-2010
>     13:34:45":1/units=seconds/t0="1-jan-2008"/cal=gregorian mytaxis
>
>     let tt = t[gt=mytaxis]
>     LET x1 = `tt,return=lsize`
>      !-> DEFINE VARIABLE x1 = 76340086
>
>
>     define symbol time_end = 02-JUN-2010 13:34:46
>
>     def
>     axis/t="1-jan-2008":"($time_end)":1/units=seconds/t0="1-jan-2008"/cal=gregorian
>     mytaxis
>      !-> def axis/t="1-jan-2008":"02-JUN-2010
>     13:34:46":1/units=seconds/t0="1-jan-2008"/cal=gregorian mytaxis
>
>     Replacing definition of axis MYTAXIS
>     let tt = t[gt=mytaxis]
>     LET x2 = `tt,return=lsize`
>      !-> DEFINE VARIABLE x2 = 76340087
Thanks Ansley,
Indeed your method works perfectly well. Your explanation and the 
limitations are clear.
That's what I call an efficient support and also why I promote ferret 
and use it every day. Almost.

Patrick

--
LSCE/IPSL, Laboratoire CEA-CNRS-UVSQ
Data Analysis and Visualization Engineer
IPSL Global Climate Modelling Group
--