How Can I Preserve Floating-Point Precision When Marshaling JSON in Go?

Preserving Floating Point Precision in JSON Marshaling

In Go, the json.Marshal() function often removes trailing zeros from floating-point numbers during JSON serialization. This can be problematic if the consuming application requires the full precision of the original value.

To overcome this issue, consider defining a custom data type that encapsulates the floating-point value. Here's an example:

type PreservedFloat float64

func (f PreservedFloat) MarshalJSON() ([]byte, error) {
    // Preserve trailing zeros in the JSON representation.
    if float64(f) == float64(int(f)) {
        return []byte(strconv.FormatFloat(float64(f), 'f', 1, 32)), nil
    }
    return []byte(strconv.FormatFloat(float64(f), 'f', -1, 32)), nil
}

In this implementation:

• The PreservedFloat type serves as a wrapper around a standard float64.
• The MarshalJSON method overrides the default JSON serialization behavior.
• If the value has an integer representation (no fractional part), it is formatted with one decimal place to preserve the trailing zero.
• Otherwise, the value is formatted without specifying a precision, allowing it to retain its full accuracy.

By using PreservedFloat, you can control the JSON representation of your floating-point values, ensuring that they retain their precision even after marshaling.

The above is the detailed content of How Can I Preserve Floating-Point Precision When Marshaling JSON in Go?. For more information, please follow other related articles on the PHP Chinese website!