| INTERPOLATE note [message #27363] |
Wed, 17 October 2001 17:07 |
Robert S. Hill
Messages: 11
Registered: January 1998
|
Junior Member |
|
|
This may have been remarked on before, but the built-in
INTERPOLATE function seems to treat its "floating-point
subscript" as single precision no matter what. I don't
see anything about this in the documentation. For
example,
p = [1.0d0,2.0d0]
x = 0.123123123123123d0
print,interpolate(p,x),format='(E25.15)'
1.123123124241829E+00
print,interpolate(p,float(x)),format='(E25.15)'
1.123123124241829E+00
Whereas, using the library routine INTERPOL,
print,interpol(p,[0.d0,1.d0],x),format='(E25.15)'
1.123123123123123E+00
print,interpol(p,[0.d0,1.d0],float(x)),format='(E25.15)'
1.123123124241829E+00
The first INTERPOL result is what I expected from both
routines. The second INTERPOL result is the same as
what INTERPOLATE gives, and it shows that
INTERPOLATE converts the "subscript" argument to
single-precision.
I personally think it would be more straightforward
if built-in routines did their default typecasting
in the same way as ordinary arithmetic expressions,
and left up to me the question of whether I really
need all those digits of precision. (In my application,
as a matter of fact I probably don't.)
- Bob
--
--------------------------------------------
Robert S. Hill 301-286-3577
Robert.S.Hill.1@gsfc.nasa.gov
Science Systems and Applications, Inc.
Code 681, NASA GSFC, Greenbelt, MD 20771
--------------------------------------------
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
