English
English
In PHP, the pack function is a very practical tool used to pack data into binary strings according to a specified format, often utilized in network transmission, file storage, and similar contexts. Particularly when dealing with floating-point numbers, pack provides two format codes: "f" and "d". Although they appear similar, there are essential differences between them. Today, let's take a closer look at these two and the floating-point precision traps you might have stepped into.
The basic usage of the pack function is:
pack(string $format, mixed ...$values): string
It converts the subsequent arguments into binary data based on the format string.
Among them, common floating-point format codes include:
"f": single-precision floating-point number (float, 4 bytes)
"d": double-precision floating-point number (double, 8 bytes)
"f" corresponds to a 4-byte single-precision floating-point number, that is, the float type, following the IEEE 754 32-bit floating-point format.
"d" corresponds to an 8-byte double-precision floating-point number, that is, the double type, following the IEEE 754 64-bit floating-point format.
This means "d" can represent a wider range and higher precision of floating-point numbers, while "f" takes less storage space but with limited precision.
Single-precision floating-point numbers ("f") have approximately 7 significant digits of precision.
Double-precision floating-point numbers ("d") have about 15-16 significant digits of precision.
If you need highly accurate floating-point data, it is recommended to use "d".
The floating-point formats in pack are affected by the machine's byte order (endianness). Although most systems use little-endian, special care is needed during cross-platform transmission.
<?php
$value = 0.1 + 0.2;
<p>$packed_f = pack("f", $value);<br>
$packed_d = pack("d", $value);</p>
<p>echo "Original value: $value\n";<br>
echo "Single-precision packed hex: ".bin2hex($packed_f)."\n";<br>
echo "Double-precision packed hex: ".bin2hex($packed_d)."\n";<br>
Sample output:
Original value: 0.3
Single-precision packed hex: 9a99993e
Double-precision packed hex: 9a9999999999b93f
You can see that although the input is 0.3, the binary representations for single and double precision differ. This is because floating-point numbers cannot precisely represent some decimal fractions, and single precision has larger errors.
When storing or transmitting high-precision floating-point numbers, prefer "d".
Use "f" to save space when precision requirements are not high.
When dealing with network protocols or file formats, clearly follow the floating-point format specified by the protocol to avoid misunderstandings.
For cross-platform applications, consider byte order issues and ensure consistent endianness when using pack and unpack.
Official PHP manual page for the pack function:
