Re: [ferret_users] defining a variable in a loop

[Jaison Kurian <jaison@xxxxxxxxxxxxxxxxxx>]

Hi Angélique,
              Please remember that Ferret alwyas works on the complete 
arrays, 
more  so with the definition of new variables. With "LET" command, you cannot 
specify array indices. Hence a C/fortran style procedure will not work with 
Ferret. To do the data extraction in your case, i think the SAMPLEXY external 
function is the best solution. Please find the details about SAMPLEXY function
in User Manual, Ch3 Sec2.3.34. (SAMPLEIJ is described in Ch3 Sec2.3.32).  

Assuming pt_lon/pt_lat in d=1 is a one dimentioal array along X-axis, the 
exact 
answer to what you asked is as follows:

  let ipts = pt_lon[d=2,i=1:47]
  let jpts = pt_lat[d=2,i=1:47]

  let T_model = SAMPLEIJ(votemper[d=1,l=8], ipts, jpts)
  let S_model = SAMPLEIJ(vosaline[d=1,l=8], ipts, jpts)

Now, assume that the pt_lon/pt_lat in d=2 corresponds to lon and lat of some 
observation and votemper/vosaline in d=1 corresponds to some gridded model or 
observational fields. In this case, you have to specify X and Y while 
defining 
the new variable, instead of I and J. The proper method is as follows:

  let xpts = pt_lon[d=2,i=1:47]
  let ypts = pt_lat[d=2,i=1:47]

  let T_model = SAMPLEXY(votemper[d=1,l=8], xpts, ypts)
  let S_model = SAMPLEXY(vosaline[d=1,l=8], xpts, ypts)

Please let me know, if you have any questions.

Jaison Kurian



On Tue, 06 Nov 2007 11:01:55 +0100 (CET), melet wrote
> Hi everybody,
> 
> I would like to define a variable according a direction in a loop, 
> but I can't manage to do that.
> 
> I do :
> 
> repeat/range=1:47/name=jj \
> (let i_pt=pt_lon[d=2,i=`jj`]; \
> let j_pt=pt_lat[d=2,i=`jj`]; \
> let T_model=votemper[d=1,i=`i_pt`,j=`j_pt`,l=8]; \
> let S_model=vosaline[d=1,i=`i_pt`,j=`j_pt`,l=8])
> 
> Of course, my variables "S_model" and "T_model" are overwritten at each
> step, whereas I would like those variables to be defined in the
> i-direction with i being increased in the loop.
> 
> I would like to do :
> 
> repeat/range=1:47/name=jj \
> (let i_pt=pt_lon[d=2,i=`jj`]; \
> let j_pt=pt_lat[d=2,i=`jj`]; \
> let/i=`jj` T_model=votemper[d=1,i=`i_pt`,j=`j_pt`,l=8]; \
> let/i=`jj` S_model=vosaline[d=1,i=`i_pt`,j=`j_pt`,l=8])
> 
> So that T_model and S_model would have 47 components in the i-direction.
> But that doesn't seem to be allowed. How can I do ?
> 
> Thanks,
> 
> Angélique Melet
> 
> -- 
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.


--
Centre for Atmospheric
       and
  Oceanic Sciences,
Indian Institute of Science
Bangalore-560 012
India


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.