Precision Loss in Floating-Point to Integer Casts
You'll learn how casting floating-point numbers to integers discards fractional parts and can exceed integer limits.
When you cast a double or float to an int, long, short, or byte, Java performs a narrowing conversion that fundamentally changes how the number is represented. Unlike integer-to-integer narrowing (which wraps around), floating-point-to-integer casts have two important behaviors:
The Fractional Part Vanishes
The decimal portion is simply discarded, not rounded. Think of it like sawing off the fractional part of a ruler—3.9 inches becomes 3 inches, not 4.
double price = 19.99;
int dollars = (int) price; // dollars = 19, not 20
The .99 disappears completely. This is called truncation toward zero: positive numbers round down, negative numbers round up (toward zero).
double temp = -5.8;
int rounded = (int) temp; // rounded = -5, not -6
Integer Overflow Danger
If the floating-point value exceeds the integer type's range, the result is unpredictable—often wrapping to the minimum or maximum value of that type. A double holding 1.5e10 (15 billion) cast to int won't fit, since int maxes out around 2.1 billion.
double huge = 3.5e9;
int compressed = (int) huge; // Result: 2147483647 (Integer.MAX_VALUE)
Always ensure your floating-point value fits within the target integer type's range before casting.
Key Takeaway: Casting float or double to an integer type truncates the fractional part (never rounds) and can overflow integer limits, potentially producing unexpected values.