Skip to main content

Max

Function Max( const A, B : Extended): Extended

Example​

procedure ScriptEvent(var Value: variant);
begin
  Value := Max(12.5, 19.75);
end;

Usage​

Returns the greater of two Extended values using Velox's floating-point maximum routine.

Parameters

NameTypeDescription
AExtended, constFirst value.
BExtended, constSecond value.

Returns

The selected input value. For ordinary unequal finite values, this is the numerically greater operand.

Behaviour

There is no tolerance or rounding step. Equal finite values remain equal, and infinities order normally. Because the function returns one input rather than calculating a new arithmetic result, the selected value's floating-point representation is preserved.

Errors

Finite, infinite and NaN operands have no expected exception path. A caller value that cannot be converted to Extended can fail before entry.

Usage notes

Use explicit finite-value validation when NaN must be rejected. Use CompareValue when the flow needs a three-way, tolerance-aware relationship rather than one selected operand.

Additional Technical Info​

Max returns the greater of two Extended values. Velox exposes only the Extended wrapper, even though Delphi provides several overloads.

The example returns 19.75. It is source-reviewed and is not executed by the documentation workflow.

Implementation

vxCommonNumber.Max delegates directly to the Extended overload of System.Math.Max. Current Delphi source compares the two operands, propagates a NaN operand and handles equal signed zeros specially so that positive zero wins.

Edge cases and quirks

  • If either operand is NaN, the current Delphi implementation returns that NaN. This differs from a simple if A > B then A else B expression when A is NaN.
  • Max(-0.0, +0.0) and Max(+0.0, -0.0) return positive zero.
  • Extended precision is platform-dependent. It is commonly 80-bit on Win32 and aliases Double on Win64, so conversion into the parameters can already have rounded a value.
  • This is a binary floating-point maximum, not a decimal/business rounding rule and not validation that either input is finite.

Side effects

None.

Performance and concurrency

The operation is constant-time, allocates no managed data and uses no shared mutable state.

Related entries

  • Min selects the lesser operand.
  • CompareValue returns an approximate ordering.
  • MoreThanValue returns a Boolean greater-than relationship after tolerance handling.

External references

Created 2026-07-15