Excel INT函数转Delphi实现：现有方案优化咨询

Hey there! Nice work getting the Excel INT behavior working correctly already—let's dive into better ways to implement this.

First, let's recap the key detail: Excel's INT function rounds down towards negative infinity, which is different from Delphi's built-in Trunc (which rounds towards zero). For example:

• INT(-465.9862462) returns -466
• Trunc(-465.9862462) returns -465 (not what we want for Excel compatibility)

The Optimal Solution: Use Math.Floor #

Delphi's Math unit has a Floor function that does exactly what Excel's INT does—it always rounds down to the nearest integer towards negative infinity. This is the cleanest, most efficient implementation you can use, no custom branching required.

Here's how to use it:

1. Make sure you add Math to your uses clause.
2. Replace your custom INT function with a direct call to Floor:

uses Math;

// Example usage matching your scenario
var
  OriginalValue, CalculatedValue: Double;
  ExcelIntResult: Integer;
begin
  OriginalValue := -465.9862462;
  CalculatedValue := 14.01375378; // Assuming your calculation leads here
  
  // Get Excel-style INT result for either value
  ExcelIntResult := Floor(OriginalValue); // Returns -466
  ExcelIntResult := Floor(CalculatedValue); // Returns 14
end;

Why This Is Better Than Custom Code #

• No manual branching: You avoid writing conditional logic for positive/negative values, reducing the chance of bugs.
• Optimized performance: Floor is a native RTL function, so it's already tuned for speed.
• Full compatibility: It matches Excel's INT behavior perfectly for all edge cases (like negative numbers, zero, and positive fractions).

If you were previously using a custom function like this:

function CustomExcelInt(Value: Double): Integer;
begin
  if Value >= 0 then
    Result := Trunc(Value)
  else
    Result := Trunc(Value) - 1;
end;

Replacing it with Floor is a drop-in improvement—it does the same thing but more cleanly.

内容的提问来源于stack exchange，提问作者Seti Net