Quoting constructs

From Rosetta Code
Quoting constructs is a draft programming task. It is not yet considered ready to be promoted as a complete task, for reasons that should be found in its talk page.

Pretty much every programming language has some form of quoting construct to allow embedding of data in a program, be it literal strings, numeric data or some combination thereof.

Show examples of the quoting constructs in your language. Explain where they would likely be used, what their primary use is, what limitations they have and why one might be preferred over another. Is one style interpolating and another not? Are there restrictions size of the quoted data? The type? The format?

This is intended to be open-ended and free form. If you find yourself writing more than a few thousand words of explanation, summarize and provide links to relevant documentation; but do provide at least a fairly comprehensive summary here, on this page, NOT just a link to [See the language docs].

Note: This is primarily for quoting constructs for data to be "embedded" in some way into a program. If there is some special format for external data, it may be mentioned but that isn't the focus of this task.


package main
import (
/* Quoting constructs in Go. */
// In Go a Unicode codepoint, expressed as a 32-bit integer, is referred to as a 'rune'
// but the more familiar term 'character' will be used instead here.
// Character literal (single quotes).
// Can contain any single character including an escaped character.
var (
rl1 = 'a'
rl2 = '\'' // single quote can only be included in escaped form
// Interpreted string literal (double quotes).
// A sequence of characters including escaped characters.
var (
is1 = "abc"
is2 = "\"ab\tc\"" // double quote can only be included in escaped form
// Raw string literal(single back quotes).
// Can contain any character including a 'physical' new line but excluding back quote.
// Escaped characters are interpreted literally i.e. `\n` is backslash followed by n.
// Raw strings are typically used for hard-coding pieces of text perhaps
// including single and/or double quotes without the need to escape them.
// They are particularly useful for regular expressions.
var (
rs1 = `

rs2 = `This is one way of including a ` + "`" + ` in a raw string literal.`
rs3 = `\d+` // a sequence of one or more digits in a regular expression
func main() {
fmt.Println(rl1, rl2) // prints the code point value not the character itself
fmt.Println(is1, is2)
re := regexp.MustCompile(rs3)
/* None of the above quoting constructs can deal directly with interpolation.
This is done instead using library functions.

// C-style using %d, %f, %s etc. within a 'printf' type function.
n := 3
fmt.Printf("\nThere are %d quoting constructs in Go.\n", n)
// Using a function such as fmt.Println which can take a variable
// number of arguments, of any type, and print then out separated by spaces.
s := "constructs"
fmt.Println("There are", n, "quoting", s, "in Go.")
// Using the function os.Expand which requires a mapper function to fill placeholders
// denoted by ${...} within a string.
mapper := func(placeholder string) string {
switch placeholder {
case "NUMBER":
return strconv.Itoa(n)
case "TYPES":
return s
return ""
fmt.Println(os.Expand("There are ${NUMBER} quoting ${TYPES} in Go.", mapper))
97 39
abc "ab	c"


This is one way of including a ` in a raw string literal.

There are 3 quoting constructs in Go.
There are 3 quoting constructs in Go.
There are 3 quoting constructs in Go.


Single quotes are used for single ascii characters, eg 'A'. Multibyte unicode characters are typically held as utf-8 strings.
Double quotes are used for single-line strings, with backslash interpretation, eg "one\ntwo\nthree\n".
The concatenation operator & along with a couple more quotes can certainly be used to mimic string continuation, however it is technically an implementation detail rather than part of the language specification as to whether that occurs at compile-time or run-time.
Phix does not support interpolation other than printf-style, eg printf(1,"Hello %s,\nYour account balance is %3.2f\n",{name,balance}).
Back-ticks and triple-quotes are used for multi-line strings, without backslash interpretation, eg

constant t123 = `

or (entirely equivalent, except the following can contain back-ticks which the above cannot, and vice versa for triple quotes)

constant t123 = """

Both are also equivalent to the top double-quote one-liner. Note that a single leading '\n' is automatically stripped.
Several builtins such as substitute, split, and join are often used to convert such strings into the required internal form.
Regular expressions are usually enclosed in back-ticks, specifically to avoid backslash interpretation.
You can also declare hexadecimal strings, eg

x"1 2 34 5678_AbC" -- same as {0x01, 0x02, 0x34, 0x56, 0x78, 0xAB, 0x0C}
-- note however it displays as {1,2,52,86,120,171,12}
-- whereas x"414243" displays as "ABC" (as all chars)

Literal sequences are represented with curly braces, and can be nested to any depth, eg

{2, 3, 5, 7, 11, 13, 17, 19}
{1, 2, {3, 3, 3}, 4, {5, {6}}}
{{"John", "Smith"}, 52389, 97.25}
{} -- the 0-element sequence


The Perl philosophy, which Raku thoroughly embraces, is that "There Is More Than One Way To Do It" (often abbreviated to TIMTOWDI). Quoting constructs is an area where this is enthusiastically espoused.

Raku has a whole quoting specific sub-language built in called Q. Q changes the parsing rules inside the quoting structure and allows extremely fine control over how the enclosed data is parsed. Every quoting construct in Raku is some form of a Q syntactic structure, using adverbs to fine tune the desired behavior, though many of the most commonly used have some form of "shortcut" syntax for quick and easy use. Usually, when using an adverbial form, you may omit the Q: and just use the adverb.

In general, any and all quoting structures have theoretically unlimited length, in practice, are limited by memory size, practically, it is probably better to limit them to less than a gigabyte or so, though they can be read as a supply, not needing to hold the whole thing in memory at once. They can hold multiple lines of data. How the new-line characters are treated depends entirely on any white-space adverb applied. The Q forms use some bracketing character to delimit the quoted data. Usually some Unicode bracket ( [], {}, <>, ⟪⟫, whatever,) that has an "open" and "close" bracketing character, but they may use any non-indentifier character as both opener and closer. ||, //, ??, the list goes on. The universal escape character for constructs that allow escaping is backslash "\".

The following exposition barely scratches the surface. For much more detail see the Raku documentation for quoting constructs for a comprehensive list of adverbs and examples.

The most commonly used
  • Q[ ], common shortcut: 「 」
The most basic form of quoting. No interpolation, no escape sequences. What is inside is what you get. No exceptions.
「Ze backslash characters!\ \Zay do NUSSING!! \」 -> Ze backslash characters!\ \Zay do NUSSING!! \

  • "Single quote" quoting. - Q:q[ ], adverbial: q[ ], common shortcut: ' '
No interpolation, but allow escape sequences.
E.G. 'Don\'t panic!' -> Don't panic!

  • "Double quote" quoting. - Q:qq[ ], adverbial: qq[ ], common shortcut: " "
Interpolates: embedded variables, logical characters, character codes, continuations.
E.G. "Hello $name, today is {Date.today} \c[grinning face] \n🦋" -> Hello Dave, today is 2020-03-25 😀
Where $name is a variable containing a name (one would imagine), {Date.today} is a continuation - a code block to be executed and the result inserted, \c[grinning face] is the literal emoji character 😀 as a character code, \n is a new-line character and 🦋 is an emoji butterfly. Allows escape sequences, and indeed, requires them when embedding data that looks like it may be an interpolation target but isn't.

Every adverbial form has both a q and a qq variant to give the 'single quote' or "double quote" semantics. Only the most commonly used are listed here.

  • "Quote words" - Q:qw[ ], adverbial: qw[ ], common shortcut: < >
No interpolation, but allow escape sequences. (Inherited from the q[] escape semantics
E.G. < a β 3 Б 🇩🇪 >
Parses whatever is inside as a white-space separated list of words. Returns a list with all white space removed. Any numeric values are returned as allomorphs.
That list may be operated on directly with any listy operator or it may be assigned to a variable.
say < a β 3 Б 🇩🇪 >[*-1] # What is the last item in the list? (🇩🇪)
say +< a β 3 Б 🇩🇪 > # How many items are in the list? (5)

  • "Quote words with white space protection" - Q:qww[ ], adverbial: qww[ ]
May preserve white space by enclosing it in single or double quote characters, otherwise identical to qw[ ].
say qww< a β '3 Б' 🇩🇪 >[2] # Item 2 in the list? (3 Б)

  • "Double quote words" quoting. - Q:qqw[ ], adverbial: qqw[ ], common shortcut: << >> or « »
Interpolates similar to standard double quote, but then interprets the interpolated string as a white space separated list.

  • "Double quoted words with white space protection" - Q:qqww[ ], adverbial: qqww[ ]
Same as qqw[ ] but retains quoted white space.

  • "System command" - Q:qx[ ], adverbial: qx[ ]
Execute the string inside the construct as a system command and return the result.

  • "Here doc format" - Q:q:to/END/ END, adverbial: q:to/END/; END
Return structured text between two textual delimiters. Depending on the adverb, may or may not interpolate (same rules as other adverbial forms.) Will return the text with the same indent as the indent of the final delimiter. The text delimiter is user chosen (and is typically, though not necessarily uppercase) as is the delimiter bracket character.

There are other adverbs to give precise control what interpolates or doesn't, that may be applied to any of the above constructs. See the doc page for details. There is another whole sub-genre dedicated to quoting regexes.


Quoting text: zkl has two types of text: parsed and raw. Strings are limited to one line, no continuations.

Parsed text is in double quotes ("text\n") and escape ("\n" is newline, UTF-8 ("\Ubd;" or "\u00bd"), etc).

"Raw" text is unparsed, usefull for things like regular expressions and unit testing of source code. It uses the form 0'<sigil>text<sigil>. For example 0'<text\n> is the text "text\\n". There is no restriction on sigil (other than it is one character).

Text blocks are nultiple lines of text that are gathered into one line and then evaluated (thus can be anything, such as string or code and are often mixed). #<<< (at the start of a line) begins and ends the block. A #<<<" beginning tag prepends a " to the block. For example:


is parsed as text:="\nA\n";

Other data types are pretty much as in other languages.