String Functions

 

    Currently there are three functions for manipulating strings: paste, spaste, and split. (Clearly more are needed.)

paste     takes a list of values, converts them all to scalar string's, and returns their concatenation as a scalar string value. For example,

    a := [2,3,5]
    paste( "the first three primes are", a )

yields

    the first three primes are [2 3 5]

The []'s seen here in the string representation of the vector a only occur for a numeric value with more than one element.

Similarly,

    paste( "hello", "there" )

is equivalent to the string constant

    'hello there'

By default, the string values are concatenated together using a   single space. The optional sep= argument can be used to specify a string to use instead. For example,

    paste("hello", "there", "how", "are", "you?",
          sep="XYZ")

yields

    helloXYZthereXYZhowXYZareXYZyou?

Note that the arguments to paste are first converted to scalar string's, and then concatenated together. So

    paste( "hello there", 1:3, sep="" )

yields

    hello there[1 2 3]

and not

    hellothere[123]

If a single string is passed to paste, the elements of the string are concatenated. For example:

    paste("a b c d e", sep='' )

results in the string "abcde". With multiple strings, the individual strings are concatenated, but not the individual elements of the strings.

spaste     is simply a version of paste with the separator set to an empty string. It is defined using:

    func spaste(...) paste(...,sep='')

This form of paste is common enough that it merits its own simple form.

split     is basically the inverse of paste. It takes a single argument, converts it to a scalar string, and splits it into words at each block of whitespace, just as string constants are constructed when enclosed in double-quotes (see § 3.3.1, page ). Thus

    split('hello there how are you?')

is equivalent to

    "hello there how are you?"

that is, it yields a five-element string vector.

You can also call   split with a second argument, giving a string of characters at which it should break the string. For example,  

    split("hello   there how are you", "eo")

yields the equivalent of

    ['h', 'll', ' th', 'r', ' h', 'w ar', ' y', 'u']

Here the first element is 'h', the second is 'll', the third ' th', and so forth. The presence of the single leading space in ' th' may be surprising. What happened is that first split converted

    "hello   there how are you"

to a scalar value, equivalent to

    'hello there how are you'

since when the double-quoted constant was constructed all information about the number of blanks between words was lost. Next split broke the scalar into words at every occurrence of an 'e' or an 'o', but not at each blank like it would without the second argument. If a blank had been included in the second argument then these extra blanks naturally disappear:

    split("hello   there how are you", 'eo ')

yields the equivalent of

    "h ll th r h w ar y u"

Note that you have to enclose the second argument in single-quotes, otherwise the blank would have been removed.

If the second argument is a null string, the first string is broken up into individual characters. Here is an example:

    split("how are you", '')

yields the string:

    "h o w   a r e   y o u"

note that the fourth and eighth elements are space characters.