FLOAT

_FLOAT variables are 80-bits floating point numerical values up to nineteen digits precision (see IEEE 754 Extended precision). This variable type is new to QB64(PE), it is not available in QuickBASIC/QBasic.

Syntax

Explicit declaration

DIM variable AS _FLOAT 'AS type variant

DIM variable## 'type suffix variant (##)

Implicit for a specific letter range

_DEFINE F-G AS _FLOAT 'all variables beginning with F(f) or G(g)

Implicit without declaration

variable## 'simply always use the type suffix

Note

• Although the numbers are 80-bits wide (10 bytes), QB64(PE) always allocates 32 bytes to store this value. Hence, using LEN on any _FLOAT variable will always return 32 and the _MK$ function returns a 32 chars long string when converting a _FLOAT number.

Description

• Values can range up to 19 digits. Decimal point accuracy depends on whole value places taken.
• The suffix ## can also be placed after a literal numerical value to designate the value as _FLOAT.
• Values may be given as real number (123.456) or scientific notation using F as exponent marker (1.23456F+2).
• PRINT usually takes the shorter of both notations (depending on the value) to output the numbers. In alternative you may use PRINT USING to designate your desired output format.
• Floating decimal point numerical values cannot be _UNSIGNED.
• Values can be converted to 32-bytes STRING values using _MK$(_FLOAT, value) and back to numbers with _CV(_FLOAT, value$).

Some useful constants are available since QB64-PE v4.0.0

• _SIZE_OF_FLOAT = 32 'in bytes
• _FLOAT_MIN## = -1.189731495357231765F+4932 'smallest normal number
• _FLOAT_MAX## = 1.189731495357231765F+4932 'largest normal number
• _FLOAT_MIN_FRAC## = 3.362103143112093506F-4932 'smallest normal number closest to zero

Availability

• 

  all

• 

  all

• 
• 

  yes

• 

  yes

• 

  yes

• In all QB64(PE) versions there's the problem, that STR$ can't correctly handle values exceeding the DOUBLE range. This unfortunatly also affects the PRINT output as well as INPUT, INPUT # and READing from DATA.
  • We'll see if we can fix these issues in a future version in such a way, that it don't breaks QB4.5 compatiblity.
• By now, you may work around the input/data issues by providing/reading _FLOAT values as number strings and pass them through VAL afterwards to convert it into numerics.
• Since QB64-PE v4.0.0 there's also the _TOSTR$ function, which can be used to convert full range _FLOAT values into number strings. You can then simply PRINT the number strings.

See also

• SINGLE, DOUBLE
• DIM, _DEFINE
• _MK$, _CV
• LEN, Constants
• Variable Types

---

Navigation: Main Page with Articles and Tutorials Keyword Reference - Alphabetical Keyword Reference - By usage Report a broken link