MaplePrimes - answers and comments on Question, Index a Table with a float?

rationals

Joe Riel — Tue, 16 Feb 2010 07:48:43 Z

You could avoid the issue by using rationals, say 1/10 or 1/20.

floats

Robert Israel — Tue, 16 Feb 2010 12:23:48 Z

As your examples show, it's generally a bad idea to use floats to index a table, because of this issue (and others involving roundoff error).
But you might use this procedure to remove trailing zeros from a float:

> RemoveTrailingZeros:= proc(x) local S; S:= sprintf("%s0.%dg","%",Digits); parse(sprintf(S,x)); end proc;

> RemoveTrailingZeros(0.20);

0.2

Thanks! I use

ihayles — Tue, 16 Feb 2010 22:58:08 Z

Thanks! I use convert(...,rational) frequently, but removing zeros is a good idea, too. I wonder what Maple is doing internally, since evalb(0.2=0.20) returns true....evalb([0.2]=[0.20]) is false, so as the index and table gets built, equality is harder to come by...

Equality

Joe Riel — Tue, 16 Feb 2010 23:20:29 Z

For numeric quantities, Maple compares the numeric values. For other quantities Maple generally compares addresses. So two lists of numerics are equal only if they are the same list.