Compression of numeric strings

algorithm, compression

Solution

Assuming you can have floating point numbers, you have a possibility of 11 symbols:

`[0,1,2,3,4,5,6,7,8,9, .]`

This means that you need 4 bits per symbol. 3 bits can only represent a maximum of 8 symbols. You can easily use 4 bits per each symbol and get a lot of compression.

If you only have integer digits in your string, an easy solution is to convert to hexidecimal and you can use 4 bits per symbol still while getting a better compression ratio. (since there are no wasted bits with 16 symbols)

If you use Huffman compression you will get an optimal bits/per symbol ratio. You can read more about Huffman compression here.

Problem

Original source

Related problems

• C or C++ BigInt library on Microsoft Windows