Newer versions of matlab have mod implemented for int64, so maybe it is time to upgrade.
a = intmax('int64'); %for example
%use static method 'ValueOf' of bigInteger to convert int64 (long) into a BigInteger:
bi = javaMethod('valueOf', 'java.math.BigInteger', a)
%modulo operation needs another BigInteger
m = bi.mod(javaMethod('valueOf', 'java.math.BigInteger', 48));
%unfortunately matlab converts the long value return by the 'longValue' method into a double
%so the following may result in loss of precision
result = int64(m.longValue)
%to guarantee no precision loss, go throught the byte array representation:
barr = m.toByteArray;
[~, ~, endianness] = computer; %note that toByteArray always return bytes in big endian
if endianness == 'L' %so if machine is little endian
barr = flipud(barr); %flip the byte
end
if numel(barr) < 8 %toByteArray only return as many bytes as necessary
barr(8, 1) = 0; %so expand to the 8 required for int64
end
result = typecast(int8(barr), 'int64')
As per the comment above, the simplest way to get an int64 (long) value out of a BigInteger is through the longValue method. Unfortunately, matlab converts the long that it returns into a double, resulting in a loss of precision for large integers. The workaround is to get the byte representation and type cast the byte array back to int64.
Edit: Actually if you're on Windows, a much simpler way is to go through .Net, the Math.DivRem method:
System.Math.DivRem(intmax('int64'), 48)
