Type Conversions

Values of one type can be converted to a value of another type. For example, the integer 4711 can be converted to the string "4711". (These two may look similar, but they are actually entirely different things. For example, 4711 + 3 will give you 4714, "4711" + "3" will give you "47113".)

Some conversions are done automatically. For example, in the expression "hello" + 17 the integer 17 will be automatically converted to the string "17", before the two strings are concatenated, giving the result "hello17". But some conversions need to be performed with a special operation, called an explicit type conversion, value cast, or most often simply cast. There is a special cast operator, which is written as a data type between parentheses:

(float)14    // Gives the float 14.0
(int)6.9    // Gives the int 6
(string)6.7    // Gives the string "6.7"
(float)"6.7"    // Gives the float 6.7

Here is a list of some useful casts:

Casting from	to	gives the result
int	float	The floating-point number with the same value as the integer
float	int	An integer, with the fractional part (“the decimals”) just skipped
int	string	Converts the integer to a string with a number in normal base-10 notation
float	string	Converts the floating-point number to a string
string	int	Converts the first number in the string to an integer. The number should be an integer written in normal base-10 notation.
string	float	Converts the first number, with or without decimals, in the string to a floating-point number
string	program	If the string contains a filename, compile the file and return the program. The results are cached.
string	object	First casts the string to a program as above, and then clones the result. The results are cached.
array	array(type)	This recursively casts all values in the array to type.
multiset	array	The same as doing `indices(multiset)`.

Sometimes you know more than what Pike does about data types, and another use for casts is to tell Pike what you know. For example, lets say that you have a variable with the type mixed, which means that it can contain any type of data. Lets also say that you happen to know that that variable contains an integer. If you want to use this value where an integer is required, Pike might complain about it:

mixed m;
int i;
m = 18;
i = m;    // This makes Pike suspicious if you turn on strict type checking

You can use a type cast (aka soft cast) to tell Pike that m is an integer and nothing else:

i = [int]m; // Don't worry. I promise it's an int!