Best way to embed data in source code?

I have a program that reads about 10K of data. This data will never change. It's all constants, etc.

So it seems I have these options:

• I could just include a ".dat" or ".txt" file with the program, in either binary or text format and have the program read and parse as part of its startup.
• I could unroll it all in the source code and have a subroutine that is simply thousands of "somevar[14] = 1.2;", etc. assignment lines. (Obviously, I could collapse arrays down to a single assignment, but you get my idea).
• Or...is there a way to embed the data in the source some other way? I'm thinking of a mechanism like perl's __DATA__.

I suspect it's either external .dat/.txt or lots of assignment lines, but I thought I'd ask if there are other techniques people use...?

Well it depends what format the data is in (how many data types).

If they're all ints say, and separated by commas, then this works rather neatly.

Code:

---

int mydata[] = {
#include "data.h"
};

---

I would make a small throw-away program that reads the data and spits out a comma delimited list. Perhaps using the most physically compact form of constants such as hex and choose large integers like 32 bit or 64 bit. This "source" code will then be hard-coded in your program as a global assignment.

If you're using Windows, you could embed the files as resources.

Otherwise, do what is suggested above.