Fixed width output format for numbers
#include <pcp/pmapi.h>
const char *pmNumberStr(double value);
char *pmNumberStr_r(double value, char *buf, int buflen);
cc ... -lpcp
pmNumberStr returns the address of a 8-byte buffer that holds a null-byte terminated representation of value suitable for output with fixed width fields. The pmNumberStr_r function does the same, but stores the result in a user-supplied buffer buf of length buflen, which should have room for at least 8 bytes.
The value is scaled using multipliers in powers of ``one thousand'' (the decimal ``kilo'') and has a bias that provides greater precision for positive numbers as opposed to negative numbers.
The format depends on the sign and magnitude of value as follows ( d represents a decimal digit):
value range | format | |
---|---|---|
> 999995000000000 | inf? | |
999995000000000 - 999995000000 | ddd.ddT | |
999995000000 - 999995000 | ddd.ddG | |
999995000 - 999995 | ddd.ddM | |
999995 - 999.995 | ddd.ddK | |
999.995 - 0.005 | ddd.dd | |
0.005 - -0.005 | 0.00 | |
-0.005 - -99.95 | -dd.dd | |
-99.995 - -99995 | -dd.ddK | |
-99995 - -99995000 | -dd.ddM | |
-99995000 - -99995000000 | -dd.ddG | |
-99995000000 - -99995000000000 | -dd.ddT | |
< -99995000000000 | -inf? |
At the boundary points of the ranges, the chosen format will retain the maximum number of significant digits.
pmNumberStr returns a pointer to a static buffer and hence is not thread-safe. Multi-threaded applications should use pmNumberStr_r instead.