4.8 Conditional operator

Quite often it is necessary to decide which part of the code of a script to execute and have this decision depend on the result of a calculation or external input. For these cases a construct is helpful that allows to tell the program to do something if e.g. the result of a calculation is less then 0 and to do something else if the result was larger or equal to 0. An admittedly somewhat contrived example would be the calculation of the square root of a number, making sure the number is multiplied by -1 before calculating the root in case it's negative. Using the conditional operator this can be written as

sqrt_of_val = ( val < 0.0 ) ? sqrt( -1.0 * val ) : sqrt( val );

The conditional operator is somewhat special in that it consists of both the question mark, ?, and the colon, :. It tells the program to evaluate the expression before the question mark and then, depending on the result, execute either the expression before or after the colon. In the example it is tested if val is smaller than 0, and if this is true, the square root of val multiplied by -1 is taken, otherwise the square root of just val itself.

Beside comparison for the number on the left hand side being smaller than the number on the right hand side, several other operators for comparisons exist. Now follows a complete list of such comparison operators, which can only be applied to expressions that result in numbers, e.g. either simple numbers, numerical variables or numerical results of a calculation, but never to arrays:

==	true if left hand side is equal to the right hand side
!=	true if left and right hand side are unequal
<	true if left side is less than right side
<=	true if left side is less or equal to right side
>	true if left side is greater than right side
>=	true if left side is greater or equal to right side

Please note that using comparisons between floating point numbers may not work as you may expect them to do. This is a result of the finite precision of floating point numbers. Thus, for example, 0.1 and the result of e.g. 0.7 / 7 can have different values, in which case a comparison for equality will fail! Thus comparison for equality should in principle only be used with integer numbers.

All these comparison operators actually do a calculation, resulting in a new number. When you see something like a <= b this evaluates to 1 if the expression is true (i.e. if the value of a is less or equal to the value of b), and otherwise to 0. In the test of the conditional execution operator (i.e. everything before the question mark) it is checked if the result is a non-zero number, which then is taken as true, or if it's zero, which is interpreted as meaning false. Thus instead of an expression with a comparison operator the test part could also consist of just a simple number and what is going to be executed of the two alternatives after the question mark depends on the value of this number: if it is non-zero the first alternative is used, but if it is zero the second. Thus it would be valid to write:

exp_of_val = val ? exp( val ) : 1;

This checks if val is non-zero, in which case the exponential of val is calculated by doing a function call, otherwise, if val is 0, the correct result of 1 is written out directly instead of using a slightly slower function call.

One has to be a bit careful when using the operator for conditional execution because it has a very low precedence. After

x1 = 1.0 + val < 0 ? 5 : 9;
x2 = 1.0 + ( val < 0 ? 5 : 9 );

x1 and x2 will have different values. In the first line it is checked if val plus 1.0 is smaller than 0 and x1 is set accordingly to either 5 or 9, while in the second line it is checked if val is less than 0 and, depending on the result, either 5 or 9 is added to 1.0, resulting in 6 or 10. Also the following two assignments may result in different values, at least if val is negative:

x1 = val < 0 ? 5 : 9 + 1;
x2 = ( val < 0 ? 5 : 9 ) + 1;

The value of x1 will be either set to 5 or 10 (the sum of 9 and 1, while x2 will be set to 6 or 10. Thus it is to be recommended to use parentheses to make sure that the evaluation happens as expected except in trivial cases.

Beside comparison operators also logical operators can be used in the test part of the condition operator. They are discussed in detail in the next section, but they work very similarly to the comparison operators in that they always result in a value of either 1 or 0, standing for true or false, respectively.

Finally, the same effect as with the conditional operator can also be achieved by using an IF-ELSE construct (see below), but only within the EXPERIMENT section of an EDL script - everywhere else only the conditional operator can be used.

This document was generated by Jens Thoms Toerring on September 6, 2017 using texi2html 1.82.