anonymous_namespace{explain_access_path.cc} Namespace Reference

Functions
int	IntegerDigits (double d)
	Find the number of integer digits (i.e. More...
std::string	DecimalFormat (double d, int log_precision)
	Format 'd' as a decimal number with enough decimals to get a rounding error less than d*10^log_precision, without any trailing fractional zeros. More...
std::string	EngineeringFormat (double d)
	Format 'd' in engineering format, i.e. More...
std::string	NumFormat (double d)
	Format 'd' for "EXPLAIN FORMAT=TREE" output. More...
constexpr uint64_t	Power (uint64_t base, int power)
	Integer exponentiation. More...
std::string	NumFormat (uint64_t l)
	Format 'l' for "EXPLAIN FORM=TREE" output. More...

Variables
constexpr int	kPlainNumberLength = 6
	The maximal number of digits we use in decimal numbers (e.g. More...
constexpr int	kMantissaLength = 3
	The maximal number of digits in engineering format mantissas, e.g. More...
constexpr double	kMinNonZeroNumber = 1.0e-12
	The smallest number (absolute value) that we do not format as "0". More...
constexpr int	kLogPrecision = -2
	For decimal numbers, include enough decimals to ensure that any rounding error is less than <number>*10^kLogPrecision (i.e. More...
const double	kMinPlainFormatNumber
	The smallest number (absolute value) that we format as decimal (rather than engineering format). More...

Function Documentation

DecimalFormat()

std::string anonymous_namespace{explain_access_path.cc}::DecimalFormat	(	double	d,
		int	log_precision
	)

Format 'd' as a decimal number with enough decimals to get a rounding error less than d*10^log_precision, without any trailing fractional zeros.

EngineeringFormat()

std::string anonymous_namespace{explain_access_path.cc}::EngineeringFormat

(

double

d

)

Format 'd' in engineering format, i.e.

<mantissa>e<sign><exponent> where 1.0<=mantissa<1000.0 and exponent is a multiple of 3.

IntegerDigits()

int anonymous_namespace{explain_access_path.cc}::IntegerDigits

(

double

d

)

Find the number of integer digits (i.e.

those before the decimal point) in 'd' when represented as a decimal number.

NumFormat() [1/2]

std::string anonymous_namespace{explain_access_path.cc}::NumFormat

(

double

d

)

Format 'd' for "EXPLAIN FORMAT=TREE" output.

NumFormat() [2/2]

std::string anonymous_namespace{explain_access_path.cc}::NumFormat

(

uint64_t

l

)

Format 'l' for "EXPLAIN FORM=TREE" output.

Power()

constexpr uint64_t anonymous_namespace{explain_access_path.cc}::Power ( uint64_t  base, int  power  )

constexpr

Integer exponentiation.

Variable Documentation

kLogPrecision

constexpr int anonymous_namespace{explain_access_path.cc}::kLogPrecision = -2

constexpr

For decimal numbers, include enough decimals to ensure that any rounding error is less than <number>*10^kLogPrecision (i.e.

less than 1%).

kMantissaLength

constexpr int anonymous_namespace{explain_access_path.cc}::kMantissaLength = 3

constexpr

The maximal number of digits in engineering format mantissas, e.g.

"12.3e+6".

kMinNonZeroNumber

constexpr double anonymous_namespace{explain_access_path.cc}::kMinNonZeroNumber = 1.0e-12

constexpr

The smallest number (absolute value) that we do not format as "0".

kMinPlainFormatNumber

const double anonymous_namespace{explain_access_path.cc}::kMinPlainFormatNumber

Initial value:

=
    std::pow(10, 1 - kPlainNumberLength - kLogPrecision)

The smallest number (absolute value) that we format as decimal (rather than engineering format).

kPlainNumberLength

constexpr int anonymous_namespace{explain_access_path.cc}::kPlainNumberLength = 6

constexpr

The maximal number of digits we use in decimal numbers (e.g.

"123456" or "0.00123").