Re: [ferret_users] backquotes in qualifiers

[Jaison Kurian <jaison@xxxxxxxxxxxxxxxxxx>]

Hi Ryo,
         Such behaviour from immediate mode evaluation has been observed
in couple of other situations too (but...i don't remember them precisely). 
This is one reason why i prefer "symbols" rather than defined "variables" 
(using "let" command) with Ferret. "Symbols" behave in a consistent manner
irrespective of where it is being used. Moreover you can use symbols within
grave accents (or immediate mode evaluation), as shown in the second example
below.

1. The example from your mail, using symbol :

     set data coads_climatology
     define symbol dn = `sst,return=dsetnum`
     define axis/y=-90:90:3/unit=degree my_yax
     define grid/like=sst[d=($dn)]/y=my_yax grid2
     show grid grid2

2. Using symbols within grave accents (or immediate mode evaluation)

     let mnames = TSEQUENCE({"JAN","FEB","MAR","APR"})
     REPEAT/L=1:4:1 (;\
        define symbol ll = `L`  ;\
        say "  Current Month is `mnames[l=($ll)]`" ;\
     )    

Hope this helps,

Regards,

Jaison


On Fri, 3 Aug 2007, Ryo Furue wrote:

> Hi Ferret developers and users,
> 
> This may have been already reported, but backquoted expressions seem
> to confuse the parser:
> 
>   yes? set data coads_climatology
>   yes? let dn = `sst,return=dsetnum`
>    !-> DEFINE VARIABLE dn = 1
>   yes? define axis/y=-90:90:3/unit=degree my_yax
>   yes? define grid/like=sst[d=1]/y=my_yax grid1
>   yes? define grid/like=sst[d=`dn`]/y=my_yax grid2
>    **ERROR: command syntax: my_yax
>   yes? let sst2 = sst[d=`dn`]
>    !-> DEFINE VARIABLE sst2 = sst[d=1]
>   yes? define grid/like=sst2/y=my_yax grid3
> 
> Ferret doesn't seem to like "sst[d=`dn`]" in 
> 
>    define grid/like=sst[d=`dn`]/y=my_yax grid2
> 
> As shown above, a workaround is to define an intermediate variable.
> 
> Hope this helps,
> Ryo
> 
> 

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