Sorting a Table

[paul_a_cardinale]

Hmmm

I'm confused.  Supposedly, if you interpret the 64 bits of a double FP as a 64-bit signed integer, it will sort correctly.

But I observe this:

What am I missing?

[altenbach]

I haven't studies it in detail, but it might have to do with the fact that negative integers are in Two's complement.

[paul_a_cardinale]

In both Double, and I64, the MSB is 1 for negative, 0 for positive.

[altenbach]

I was talking about the remaining bits, which are in two-s complement for negative numbers.

For example in l8,

1 = b00000001

-1 = b11111111 (i.e. not b10000001)

[paul_a_cardinale]

---

@altenbach wrote:

I was talking about the remaining bits, which are in two-s complement for negative numbers.

For example in l8,

1 = b00000001

-1 = b11111111 (i.e. not b10000001)

---

Sure.  But I believed (and have read) that interpreting an (IEEE 754) floating point number as a (two's complement) signed integer would result in the same ordering as interpreting the floating point value.

[paul_a_cardinale]

OK, I'll take another stab at this.  I fixed the numeric sorting, and added a new mode (Hierarchal).

It occurred to me that there may be a need to sort things such as paragraph identifiers that typically use hierarchal numbering: "1.1" < "1.10".

However instead of using a "Mode" input, I made a polymorphic VI.