You won't crash when you use double, and you won't lose too much; Single and long will collapse; The digits are different.
I think I should be so sure:
-Integer.
1) OK but 1k with integer;
2) 1k, use Long.
-Decimal number.
1) generally uses single.
2) If you think it will be big and there will be trouble, then double it.
Look at API functions, most of them are long returns, which saves the trouble of future development; And it is more troublesome to communicate with other software.
But if you just make a nine-nine table or something, of course, it doesn't matter which result you use. In practical engineering, this is impossible.
Make a 9-9 table, maybe you can use two bytes. I haven't tried. ....
The private declaration function lcloselib "kernel32" alias "_ lclose" (byval hfile as Long) is long.
C=lcolse(b)// The value of this c is Long.
According to reason, there shouldn't be so many items to return! , some hundreds of thousands, which is great.
It takes a long time to design in order to prepare for future development. Otherwise, it will be very troublesome to change from integer to growth in the future.
Your return type has changed, VB, VC, and other software have to adapt to you, which is very troublesome; Your software has poor compatibility.
Therefore, when the api appeared, Microsoft defined it as a Long for future development. besides ...
On the contrary, the memory is cheap now, and the dragon has almost no loss. ....