This is an old revision of the document!
Formula editor is used in the following cases:
MagicPlot uses standard IEEE 754 double precision floating-point arithmetic. Double precision floating point takes 8 bytes per number and provides a relative precision of about 16 decimal digits and magnitude range from about 10-308 to about 10+308.
MagicPlot formula editor highlights expression syntax. It also marks matching brackets:
MagicPlot formula translator is generally case sensitive, i.e. you can write
sin but not
X are different variables. You can use this feature when naming
Custom Equation Fit Curve parameters.
.) or comma (
,) as decimal separator, and separate function arguments with a semicolon (
;) in the following cases:
.) only as decimal separator, and separate function arguments with a comma (
,) or a semicolon (
You can use
E for scientific notation:
You can freely insert space characters and line breaks in formula, but do not break function names, numbers, operators. You do not need to enter special characters to indicate line break.
You can see a list of all available functions and their descriptions in
Functions tab in
Set Column Formula window and in
Help on Functions window which can be opened from menu in calculator window.
MagicPlot uses functions of Java programming language library
StrictMath to evaluate
exp, etc. These functions are available from the well-known network library
netlib as a “Freely Distributable Math Library”,
fdlibm package. The same library is widely used in many scientific computing applications.
MagicPlot supports all standard trigonometric functions (
cos, etc.). All angles are always measured in radians for clarity.
You can use the following functions to convert angles units:
deg(a)— converts angles input in radians to an equivalent measure in degrees.
rad(a)— converts angles input in degrees to an equivalent measure in radians.
The predefined constants are:
pi, Pi, PI— π = 3.1416… value (the ratio of circumference of a circle to its diameter).
e— e = 2.7183… value (the base of the natural logarithms). Note: expression
e^ais evaluated as
NAN— Not-a-Number value.
Infinity— positive infinity value which may be used in some calculations. Note: write
-inffor negative infinity.
eps— machine epsilon, gives an upper bound of the relative error due to rounding in floating point arithmetic. Note:
eps = ulp(1) = 2^(-52) = 2,2204E-16. (52 is the number of bits used to store fractional part of a number.)
MagicPlot can interpret boolean logic expressions. Zero and negative values (
<=0) are interpreted as
false and positive values (
>0) are interpreted as
true similarly to C programming language.
You can use simple logical operators which are described below. Use
The basic logical function is
if(condition, a, b). If
condition argument is
true (greater than 0) it returns the second argument (
a), else returns the third argument (
if(col(A) >= 0, col(A), -col(A))— evaluates absolute value of column A (you can use
abs(col(A))for that, of course).
if(col(B) >= 0, col(B), NaN)— returns only positive values from column B. Negative values are replaced with
NaNvalue (empty cell). You can use this expression to filter negative values if you do not want to use them in future calculations. Note that “Not-a-Number returned at row #” warning can be shown for such expression.
if(col(A) > 0 & col(B) > 0, max(col(A), col(B)), NaN)
if(x >= 0, x, -x)– fit function example.
You have to be careful if you need to check equality of two values. Due to inaccuracy of computer floating-point calculations the result of evaluation is always approximate. For example, result of
sqrt(3)^2 is number
2.9999999999999996, not exactly
3. The expression
sqrt(3)^2 == 3 is
false (it returns
0). Keep in mind that for convenience MagicPlot rounds numbers when showing on the screen, so this value will be shown as
3 in table if the number of shown fractional digits in MagicPlot preferences is not big enough.
Generally, if you want to check equality of two values you need to use some equality threshold for relative difference. That is, you should compare the modulus of relative difference of two values
b with threshold
if(abs((a-b)/a) < t, …, …).
sqrt(3)^2 - 3results something about
if(abs(sqrt(3)^2 - 3) / 3 < 1e-10, …, …)— checks equality of
3with a threshold of
| ||addition|| ||equal to|
| ||subtraction|| ||not equal to|
| ||multiplication|| ||less than|
| ||division|| ||greater than|
| ||power|| ||less than or equal to|
| ||or|| ||greater than or equal to|
Operators with lower precedence value are evaluated earlier. You can use brackets to change calculation sequence.
Expression is evaluated left-to-right, excluding repeated exponentiation operator
^ operator is right-associative like in Fortran language (evaluated right-to-left; note that in general case
a^(b^c) ≠ (a^b)^c). Hence
a^b^c is evaluated as
The reason for exponentiation being right-associative is that a repeated left-associative exponentiation operation would be less useful. Multiple appearances could (and would) be rewritten with multiplication:
(a^b)^c = a^(b*c).
1 + 2 * 3returns
(1 + 2) * 3returns
-3^2is equal to
^priority is higher than that of unary minus. The result is
2^2^3is equal to
^is right-associative operator. The result is