Re: caution about negative index values (eg i=-5:5)

[Namba Takaya <takayanamba7@xxxxxxxxxxx>]

Dear Mick 

Thank you very much for a lot of your consideration on
this matter.
I did not realize that negative index has deep meaning. I
had check your demo script and I will be careful of using
the negative index.

Best Regards,

Takaya Namba

--- Mick.Spillane@noaa.gov からのメッセージ：
> Hi Namba Takaya,
> 	There is a saying in English - "You learn something
> new
> every day", and it was true for me with your example
> using the
> negative values of "i".  I had never thought to use
> that, but a
> quick test shows that it does work if the x-axis is
> defined to
> be MODULO.  However the demo below shows that one
> needs to be wary.
> 
> ! ------------- demo of the use of negative indices
> -------------
> ! with the modulo ...
> def axis/modulo/x=0:360:60/units=longitude xax1
> def grid/x=xax1 g1 ; let/title="with modulo"
> v1=x[g=g1]
> list/i=-2:2 v1
> ! VARIABLE : with modulo
> ! 180E  / -2: -180.0
> ! 120W  / -1: -120.0
> ! 60W  /  0:  -60.0
> ! 0E  /  1:  0.0
> ! 60E  /  2:  60.0  ! all is well
> 
> ! ... but if the /modulo is missing ...
> def axis/x=0:360:60/units=longitude xax2
> def grid/x=xax2 g2 ; let/title="no modulo"
> v2=x[g=g2]
> list/i=-2:2 v2
> ! VARIABLE : no modulo
> ! 0E  / 1:  0.00
> ! 60E  / 2:  60.00  ! again as expected but with
> no warning
> 
> ! However if you mix the two ...
> list/i=-2:2 v1,v2
> ! ... an error appears
> ! **ERROR: inconsistent sizes of data regions: X
> axis
> !  Variable 2 is I=1:2
> !  expression is I-2:2
> 
> ! Now suppose we make a new variable by regridding
> v2 onto the
> ! "modulo" grid g1 ...
> let/title="v2 regridded onto g1" v3=v2[g=g1]
> list/i=-2:2 v3
> ! VARIABLE : v2 regridded onto g1
> ! 180E  / -2:  ....
> ! 120W  / -1:  ....  !!! notice the missing
> values !!!
> ! 60W  /  0:  ....
> ! 0E  /  1:  0.00
> ! 60E  /  2:  60.00
> 
> ! Missing values are introduced and there is no
> warning! I suspect
> ! that in some cases the result of a calculation
> could be affected
> ! if one were not aware of what was going on behind
> the scenes.
>
!----------------------------------------------------------------
> 	I guess the message from all this is to be a bit
> cautious
> when using negative index variables. Perhaps this is
> all well
> known to those who have been long aware of negative
> indices but, for
> the "old dogs" slow to "learn new tricks" (like me),
> I thought it
> was worth mention.
> Good luck,
> Mick
> 
> |!!! Mick.Spillane@noaa.gov !!!|
> |__Room 2070 Bldg#3 NOAA/PMEL__|
> |____Phone_:_(206)526-6780_____|
> 
> 
> 

__________________________________
Let's Celebrate Together!
Yahoo! JAPAN
http://pr.mail.yahoo.co.jp/so2005/