Can you use switch in string in c language?

Answer 1

"It's not just a good idea. It's the law!

In C, switch's and strings don't mix.

Use:

if (!strcmp(myval, "S1") {...}
else if (!strcmp(myval, "S2") {...}

etc"

-- ptrussell_nc
http://www.sqaforums.com/showflat.php?Cat=0&Number=473868&Main=473575

Answer 2

In many cases we use whichever version offers the greatest readability because code that is easy to read is also easy to maintain. However, there are some caveats. Many if statements are nothing more than thinly-veiled switch statements, and in these case a switch statement would be preferred. For example:

int f (int x) {

if (x==0) return 100;

else if (x==1) return 50;

else if (x==2) return 25;

else return 0;

}

Thinly-veiled switch statements are often used in tutorial code because people are usually more familiar with if statements than switch statements. However, switch statements are arguably easier to read and thus easier to maintain:

int f (int x) { switch (x) {

case 0: return 100;

case 1: return 50;

case 2: return 25;

default: return 0;

}

}

There can also be performance benefits. Switch statements of this form are usually more efficient than the thinly-veiled switch because every control expression in the if statement must be evaluated in turn until one evaluates true. If they all evaluate false (such as when x==3 in the example), execution passes to the final else clause. But with a switch statement, the controlling expression is evaluated just once and execution immediately jumps to the appropriate label matching the evaluation of that expression. If there is no corresponding label, execution immediately passes to the default label. Either way, the switch statement becomes a constant-time operation, whereas the thinly-veiled if statement has variable-time operation.

A good compiler implementation should produce the exact same (constant-time) machine code regardless of which we use, thus the choice is simply one of readability. However we cannot rely on all compiler implementations being equal, so it pays to be explicit.

Switch statements also provide the benefit of fall-through whereas if statements do not. That is, execution will begin at the appropriate label and will continue until a break, return or goto statement is encountered. Consider:

enum letter {vowel, consonant};

letter g (char x) {

switch (tolower (x)) {

case 'a': case 'e': case 'i': case 'o': case 'u': return vowel;

default: return consonant;

}

}

Normally we wouldn't place multiple cases on the same line as it can hinder readability, but for trivial code such as this we can obviously make exceptions. To achieve the same effect with an if statement we'd use the following equivalent:

letter g (char x) {

x = tolower(x);

if (x=='a' x=='e' x=='i' x=='o' x=='u') return vowel;

else return consonant;

}

This is arguably just as readable as the switch statement, but all bar the best compiler implementations will struggle to produce efficient machine code from this. That is, the if statement requires 5 evaluations to evaluate a consonant whereas the switch statement only requires one evaluation.

A switch statement also benefits from the fact the default label (if required) need not be the last case in a switch statement, but it must always come last in an if statement. Indeed, the order of case labels is generally immaterial in a switch statement unless we need to make use of fall-through. But being able to move the default label to an appropriate fall-through point is useful in itself. In these cases, the equivalent if statement will usually require some code duplication or require complex code that is much more difficult to read, both of which add to the maintenance burden.

Of course, where performance is essential we shouldn't rely upon our assumptions. Always test the performance.