Suggestions to represent a BIG number

[d0mufasa]

Hello all,

Basically, I am on a Linux system.

I have written code to use a key of DOUBLE. I was trying to represent a large number - so -
I thought I would do it as :

<date>.<incremental-number>

So, it would be:

20100710.000001,
20100710.000002, etc.

and then to reset the <incremental-number> back to 0 at the end of the day.

So,

It works well with the Oracle DB.
It works well with the Linux OS.

but then, I started using a SOAP product defined here:
gSOAP: SOAP C++ Web Services

and then it all went to pot

When using setting up the client (using this SOAP functionality), a number that is:

20100710.000234 just comes out as 20100710.000000

Now I am trying to redo the system using a NEW key - never mind that it took over a year to
develop.

But, before doing so, I just need to make sure that the data type representation will work.

Any ideas on the best way to represent a large key?

TIA

[Cabhan]

It sounds to me like you actually have two keys: a date and an incremental ID. Therefore, you're better off using two separate pieces of data here.

For the date part, there is already a datatype for this called time_t. It gives the number of seconds since January 1, 1970 00:00 UTC. You can parse a time_t into year, month, day, etc. by using the function gmtime() (see "man gmtime" for more information).

For the incremental number, assuming you start at 0 and increase, you can try an unsigned long int. You may also be able to use an unsigned long long int. It is unlikely that in a single day you will surpass the storage of an unsigned long int (0 - 2^32). If you may exceed even these numbers, you can try something like libgmp:
The GNU MP Bignum Library