1 @c -*- mode: texinfo; coding: utf-8 -*-
2 @c This is part of the GNU Emacs Lisp Reference Manual.
3 @c Copyright (C) 1990-1995, 1998-1999, 2001-2016 Free Software
5 @c See the file elisp.texi for copying conditions.
6 @node Strings and Characters
7 @chapter Strings and Characters
9 @cindex character arrays
13 A string in Emacs Lisp is an array that contains an ordered sequence
14 of characters. Strings are used as names of symbols, buffers, and
15 files; to send messages to users; to hold text being copied between
16 buffers; and for many other purposes. Because strings are so important,
17 Emacs Lisp has many functions expressly for manipulating them. Emacs
18 Lisp programs use strings more often than individual characters.
20 @xref{Strings of Events}, for special considerations for strings of
21 keyboard character events.
24 * Basics: String Basics. Basic properties of strings and characters.
25 * Predicates for Strings:: Testing whether an object is a string or char.
26 * Creating Strings:: Functions to allocate new strings.
27 * Modifying Strings:: Altering the contents of an existing string.
28 * Text Comparison:: Comparing characters or strings.
29 * String Conversion:: Converting to and from characters and strings.
30 * Formatting Strings:: @code{format}: Emacs's analogue of @code{printf}.
31 * Case Conversion:: Case conversion functions.
32 * Case Tables:: Customizing case conversion.
36 @section String and Character Basics
38 A character is a Lisp object which represents a single character of
39 text. In Emacs Lisp, characters are simply integers; whether an
40 integer is a character or not is determined only by how it is used.
41 @xref{Character Codes}, for details about character representation in
44 A string is a fixed sequence of characters. It is a type of
45 sequence called a @dfn{array}, meaning that its length is fixed and
46 cannot be altered once it is created (@pxref{Sequences Arrays
47 Vectors}). Unlike in C, Emacs Lisp strings are @emph{not} terminated
48 by a distinguished character code.
50 Since strings are arrays, and therefore sequences as well, you can
51 operate on them with the general array and sequence functions documented
52 in @ref{Sequences Arrays Vectors}. For example, you can access or
53 change individual characters in a string using the functions @code{aref}
54 and @code{aset} (@pxref{Array Functions}). However, note that
55 @code{length} should @emph{not} be used for computing the width of a
56 string on display; use @code{string-width} (@pxref{Size of Displayed
59 There are two text representations for non-@acronym{ASCII}
60 characters in Emacs strings (and in buffers): unibyte and multibyte.
61 For most Lisp programming, you don't need to be concerned with these
62 two representations. @xref{Text Representations}, for details.
64 Sometimes key sequences are represented as unibyte strings. When a
65 unibyte string is a key sequence, string elements in the range 128 to
66 255 represent meta characters (which are large integers) rather than
67 character codes in the range 128 to 255. Strings cannot hold
68 characters that have the hyper, super or alt modifiers; they can hold
69 @acronym{ASCII} control characters, but no other control characters.
70 They do not distinguish case in @acronym{ASCII} control characters.
71 If you want to store such characters in a sequence, such as a key
72 sequence, you must use a vector instead of a string. @xref{Character
73 Type}, for more information about keyboard input characters.
75 Strings are useful for holding regular expressions. You can also
76 match regular expressions against strings with @code{string-match}
77 (@pxref{Regexp Search}). The functions @code{match-string}
78 (@pxref{Simple Match Data}) and @code{replace-match} (@pxref{Replacing
79 Match}) are useful for decomposing and modifying strings after
80 matching regular expressions against them.
82 Like a buffer, a string can contain text properties for the characters
83 in it, as well as the characters themselves. @xref{Text Properties}.
84 All the Lisp primitives that copy text from strings to buffers or other
85 strings also copy the properties of the characters being copied.
87 @xref{Text}, for information about functions that display strings or
88 copy them into buffers. @xref{Character Type}, and @ref{String Type},
89 for information about the syntax of characters and strings.
90 @xref{Non-ASCII Characters}, for functions to convert between text
91 representations and to encode and decode character codes.
93 @node Predicates for Strings
94 @section Predicates for Strings
95 @cindex predicates for strings
96 @cindex string predicates
98 For more information about general sequence and array predicates,
99 see @ref{Sequences Arrays Vectors}, and @ref{Arrays}.
101 @defun stringp object
102 This function returns @code{t} if @var{object} is a string, @code{nil}
106 @defun string-or-null-p object
107 This function returns @code{t} if @var{object} is a string or
108 @code{nil}. It returns @code{nil} otherwise.
111 @defun char-or-string-p object
112 This function returns @code{t} if @var{object} is a string or a
113 character (i.e., an integer), @code{nil} otherwise.
116 @node Creating Strings
117 @section Creating Strings
118 @cindex creating strings
119 @cindex string creation
121 The following functions create strings, either from scratch, or by
122 putting strings together, or by taking them apart.
124 @defun make-string count character
125 This function returns a string made up of @var{count} repetitions of
126 @var{character}. If @var{count} is negative, an error is signaled.
135 Other functions to compare with this one include @code{make-vector}
136 (@pxref{Vectors}) and @code{make-list} (@pxref{Building Lists}).
139 @defun string &rest characters
140 This returns a string containing the characters @var{characters}.
148 @defun substring string &optional start end
149 This function returns a new string which consists of those characters
150 from @var{string} in the range from (and including) the character at the
151 index @var{start} up to (but excluding) the character at the index
152 @var{end}. The first character is at index zero. With one argument,
153 this function just copies @var{string}.
157 (substring "abcdefg" 0 3)
163 In the above example, the index for @samp{a} is 0, the index for
164 @samp{b} is 1, and the index for @samp{c} is 2. The index 3---which
165 is the fourth character in the string---marks the character position
166 up to which the substring is copied. Thus, @samp{abc} is copied from
167 the string @code{"abcdefg"}.
169 A negative number counts from the end of the string, so that @minus{}1
170 signifies the index of the last character of the string. For example:
174 (substring "abcdefg" -3 -1)
180 In this example, the index for @samp{e} is @minus{}3, the index for
181 @samp{f} is @minus{}2, and the index for @samp{g} is @minus{}1.
182 Therefore, @samp{e} and @samp{f} are included, and @samp{g} is excluded.
184 When @code{nil} is used for @var{end}, it stands for the length of the
189 (substring "abcdefg" -3 nil)
194 Omitting the argument @var{end} is equivalent to specifying @code{nil}.
195 It follows that @code{(substring @var{string} 0)} returns a copy of all
200 (substring "abcdefg" 0)
206 But we recommend @code{copy-sequence} for this purpose (@pxref{Sequence
209 If the characters copied from @var{string} have text properties, the
210 properties are copied into the new string also. @xref{Text Properties}.
212 @code{substring} also accepts a vector for the first argument.
216 (substring [a b (c) "d"] 1 3)
220 A @code{wrong-type-argument} error is signaled if @var{start} is not
221 an integer or if @var{end} is neither an integer nor @code{nil}. An
222 @code{args-out-of-range} error is signaled if @var{start} indicates a
223 character following @var{end}, or if either integer is out of range
226 Contrast this function with @code{buffer-substring} (@pxref{Buffer
227 Contents}), which returns a string containing a portion of the text in
228 the current buffer. The beginning of a string is at index 0, but the
229 beginning of a buffer is at index 1.
232 @defun substring-no-properties string &optional start end
233 This works like @code{substring} but discards all text properties from
234 the value. Also, @var{start} may be omitted or @code{nil}, which is
235 equivalent to 0. Thus, @w{@code{(substring-no-properties
236 @var{string})}} returns a copy of @var{string}, with all text
240 @defun concat &rest sequences
241 @cindex copying strings
242 @cindex concatenating strings
243 This function returns a new string consisting of the characters in the
244 arguments passed to it (along with their text properties, if any). The
245 arguments may be strings, lists of numbers, or vectors of numbers; they
246 are not themselves changed. If @code{concat} receives no arguments, it
247 returns an empty string.
250 (concat "abc" "-def")
252 (concat "abc" (list 120 121) [122])
254 ;; @r{@code{nil} is an empty sequence.}
255 (concat "abc" nil "-def")
257 (concat "The " "quick brown " "fox.")
258 @result{} "The quick brown fox."
264 This function always constructs a new string that is not @code{eq} to
265 any existing string, except when the result is the empty string (to
266 save space, Emacs makes only one empty multibyte string).
268 For information about other concatenation functions, see the
269 description of @code{mapconcat} in @ref{Mapping Functions},
270 @code{vconcat} in @ref{Vector Functions}, and @code{append} in @ref{Building
271 Lists}. For concatenating individual command-line arguments into a
272 string to be used as a shell command, see @ref{Shell Arguments,
273 combine-and-quote-strings}.
276 @defun split-string string &optional separators omit-nulls trim
277 This function splits @var{string} into substrings based on the regular
278 expression @var{separators} (@pxref{Regular Expressions}). Each match
279 for @var{separators} defines a splitting point; the substrings between
280 splitting points are made into a list, which is returned.
282 If @var{omit-nulls} is @code{nil} (or omitted), the result contains
283 null strings whenever there are two consecutive matches for
284 @var{separators}, or a match is adjacent to the beginning or end of
285 @var{string}. If @var{omit-nulls} is @code{t}, these null strings are
286 omitted from the result.
288 If @var{separators} is @code{nil} (or omitted), the default is the
289 value of @code{split-string-default-separators}.
291 As a special case, when @var{separators} is @code{nil} (or omitted),
292 null strings are always omitted from the result. Thus:
295 (split-string " two words ")
296 @result{} ("two" "words")
299 The result is not @code{("" "two" "words" "")}, which would rarely be
300 useful. If you need such a result, use an explicit value for
304 (split-string " two words "
305 split-string-default-separators)
306 @result{} ("" "two" "words" "")
312 (split-string "Soup is good food" "o")
313 @result{} ("S" "up is g" "" "d f" "" "d")
314 (split-string "Soup is good food" "o" t)
315 @result{} ("S" "up is g" "d f" "d")
316 (split-string "Soup is good food" "o+")
317 @result{} ("S" "up is g" "d f" "d")
320 Empty matches do count, except that @code{split-string} will not look
321 for a final empty match when it already reached the end of the string
322 using a non-empty match or when @var{string} is empty:
325 (split-string "aooob" "o*")
326 @result{} ("" "a" "" "b" "")
327 (split-string "ooaboo" "o*")
328 @result{} ("" "" "a" "b" "")
333 However, when @var{separators} can match the empty string,
334 @var{omit-nulls} is usually @code{t}, so that the subtleties in the
335 three previous examples are rarely relevant:
338 (split-string "Soup is good food" "o*" t)
339 @result{} ("S" "u" "p" " " "i" "s" " " "g" "d" " " "f" "d")
340 (split-string "Nice doggy!" "" t)
341 @result{} ("N" "i" "c" "e" " " "d" "o" "g" "g" "y" "!")
342 (split-string "" "" t)
346 Somewhat odd, but predictable, behavior can occur for certain
347 ``non-greedy'' values of @var{separators} that can prefer empty
348 matches over non-empty matches. Again, such values rarely occur in
352 (split-string "ooo" "o*" t)
354 (split-string "ooo" "\\|o+" t)
355 @result{} ("o" "o" "o")
358 If the optional argument @var{trim} is non-@code{nil}, it should be a
359 regular expression to match text to trim from the beginning and end of
360 each substring. If trimming makes the substring empty, it is treated
363 If you need to split a string into a list of individual command-line
364 arguments suitable for @code{call-process} or @code{start-process},
365 see @ref{Shell Arguments, split-string-and-unquote}.
368 @defvar split-string-default-separators
369 The default value of @var{separators} for @code{split-string}. Its
370 usual value is @w{@code{"[ \f\t\n\r\v]+"}}.
373 @node Modifying Strings
374 @section Modifying Strings
375 @cindex modifying strings
376 @cindex string modification
378 The most basic way to alter the contents of an existing string is with
379 @code{aset} (@pxref{Array Functions}). @code{(aset @var{string}
380 @var{idx} @var{char})} stores @var{char} into @var{string} at index
381 @var{idx}. Each character occupies one or more bytes, and if @var{char}
382 needs a different number of bytes from the character already present at
383 that index, @code{aset} signals an error.
385 A more powerful function is @code{store-substring}:
387 @defun store-substring string idx obj
388 This function alters part of the contents of the string @var{string}, by
389 storing @var{obj} starting at index @var{idx}. The argument @var{obj}
390 may be either a character or a (smaller) string.
392 Since it is impossible to change the length of an existing string, it is
393 an error if @var{obj} doesn't fit within @var{string}'s actual length,
394 or if any new character requires a different number of bytes from the
395 character currently present at that point in @var{string}.
398 To clear out a string that contained a password, use
401 @defun clear-string string
402 This makes @var{string} a unibyte string and clears its contents to
403 zeros. It may also change @var{string}'s length.
407 @node Text Comparison
408 @section Comparison of Characters and Strings
409 @cindex string equality
410 @cindex text comparison
412 @defun char-equal character1 character2
413 This function returns @code{t} if the arguments represent the same
414 character, @code{nil} otherwise. This function ignores differences
415 in case if @code{case-fold-search} is non-@code{nil}.
420 (let ((case-fold-search nil))
426 @defun string= string1 string2
427 This function returns @code{t} if the characters of the two strings
428 match exactly. Symbols are also allowed as arguments, in which case
429 the symbol names are used. Case is always significant, regardless of
430 @code{case-fold-search}.
432 This function is equivalent to @code{equal} for comparing two strings
433 (@pxref{Equality Predicates}). In particular, the text properties of
434 the two strings are ignored; use @code{equal-including-properties} if
435 you need to distinguish between strings that differ only in their text
436 properties. However, unlike @code{equal}, if either argument is not a
437 string or symbol, @code{string=} signals an error.
440 (string= "abc" "abc")
442 (string= "abc" "ABC")
448 For technical reasons, a unibyte and a multibyte string are
449 @code{equal} if and only if they contain the same sequence of
450 character codes and all these codes are either in the range 0 through
451 127 (@acronym{ASCII}) or 160 through 255 (@code{eight-bit-graphic}).
452 However, when a unibyte string is converted to a multibyte string, all
453 characters with codes in the range 160 through 255 are converted to
454 characters with higher codes, whereas @acronym{ASCII} characters
455 remain unchanged. Thus, a unibyte string and its conversion to
456 multibyte are only @code{equal} if the string is all @acronym{ASCII}.
457 Character codes 160 through 255 are not entirely proper in multibyte
458 text, even though they can occur. As a consequence, the situation
459 where a unibyte and a multibyte string are @code{equal} without both
460 being all @acronym{ASCII} is a technical oddity that very few Emacs
461 Lisp programmers ever get confronted with. @xref{Text
465 @defun string-equal string1 string2
466 @code{string-equal} is another name for @code{string=}.
469 @cindex locale-dependent string equivalence
470 @defun string-collate-equalp string1 string2 &optional locale ignore-case
471 This function returns @code{t} if @var{string1} and @var{string2} are
472 equal with respect to collation rules. A collation rule is not only
473 determined by the lexicographic order of the characters contained in
474 @var{string1} and @var{string2}, but also further rules about
475 relations between these characters. Usually, it is defined by the
476 @var{locale} environment Emacs is running with.
478 For example, characters with different coding points but
479 the same meaning might be considered as equal, like different grave
480 accent Unicode characters:
484 (string-collate-equalp (string ?\uFF40) (string ?\u1FEF))
489 The optional argument @var{locale}, a string, overrides the setting of
490 your current locale identifier for collation. The value is system
491 dependent; a @var{locale} @code{"en_US.UTF-8"} is applicable on POSIX
492 systems, while it would be, e.g., @code{"enu_USA.1252"} on MS-Windows
495 If @var{ignore-case} is non-@code{nil}, characters are converted to lower-case
496 before comparing them.
498 @vindex w32-collate-ignore-punctuation
499 To emulate Unicode-compliant collation on MS-Windows systems,
500 bind @code{w32-collate-ignore-punctuation} to a non-@code{nil} value, since
501 the codeset part of the locale cannot be @code{"UTF-8"} on MS-Windows.
503 If your system does not support a locale environment, this function
504 behaves like @code{string-equal}.
506 Do @emph{not} use this function to compare file names for equality, only
510 @defun string-prefix-p string1 string2 &optional ignore-case
511 This function returns non-@code{nil} if @var{string1} is a prefix of
512 @var{string2}; i.e., if @var{string2} starts with @var{string1}. If
513 the optional argument @var{ignore-case} is non-@code{nil}, the
514 comparison ignores case differences.
517 @defun string-suffix-p suffix string &optional ignore-case
518 This function returns non-@code{nil} if @var{suffix} is a suffix of
519 @var{string}; i.e., if @var{string} ends with @var{suffix}. If the
520 optional argument @var{ignore-case} is non-@code{nil}, the comparison
521 ignores case differences.
524 @cindex lexical comparison of strings
525 @defun string< string1 string2
526 @c (findex string< causes problems for permuted index!!)
527 This function compares two strings a character at a time. It
528 scans both the strings at the same time to find the first pair of corresponding
529 characters that do not match. If the lesser character of these two is
530 the character from @var{string1}, then @var{string1} is less, and this
531 function returns @code{t}. If the lesser character is the one from
532 @var{string2}, then @var{string1} is greater, and this function returns
533 @code{nil}. If the two strings match entirely, the value is @code{nil}.
535 Pairs of characters are compared according to their character codes.
536 Keep in mind that lower case letters have higher numeric values in the
537 @acronym{ASCII} character set than their upper case counterparts; digits and
538 many punctuation characters have a lower numeric value than upper case
539 letters. An @acronym{ASCII} character is less than any non-@acronym{ASCII}
540 character; a unibyte non-@acronym{ASCII} character is always less than any
541 multibyte non-@acronym{ASCII} character (@pxref{Text Representations}).
545 (string< "abc" "abd")
547 (string< "abd" "abc")
549 (string< "123" "abc")
554 When the strings have different lengths, and they match up to the
555 length of @var{string1}, then the result is @code{t}. If they match up
556 to the length of @var{string2}, the result is @code{nil}. A string of
557 no characters is less than any other string.
574 Symbols are also allowed as arguments, in which case their print names
578 @defun string-lessp string1 string2
579 @code{string-lessp} is another name for @code{string<}.
582 @defun string-greaterp string1 string2
583 This function returns the result of comparing @var{string1} and
584 @var{string2} in the opposite order, i.e., it is equivalent to calling
585 @code{(string-lessp @var{string2} @var{string1})}.
588 @cindex locale-dependent string comparison
589 @defun string-collate-lessp string1 string2 &optional locale ignore-case
590 This function returns @code{t} if @var{string1} is less than
591 @var{string2} in collation order. A collation order is not only
592 determined by the lexicographic order of the characters contained in
593 @var{string1} and @var{string2}, but also further rules about
594 relations between these characters. Usually, it is defined by the
595 @var{locale} environment Emacs is running with.
597 For example, punctuation and whitespace characters might be ignored
598 for sorting (@pxref{Sequence Functions}):
602 (sort '("11" "12" "1 1" "1 2" "1.1" "1.2") 'string-collate-lessp)
603 @result{} ("11" "1 1" "1.1" "12" "1 2" "1.2")
607 This behavior is system-dependent; e.g., punctuation and whitespace
608 are never ignored on Cygwin, regardless of locale.
610 The optional argument @var{locale}, a string, overrides the setting of
611 your current locale identifier for collation. The value is system
612 dependent; a @var{locale} @code{"en_US.UTF-8"} is applicable on POSIX
613 systems, while it would be, e.g., @code{"enu_USA.1252"} on MS-Windows
614 systems. The @var{locale} value of @code{"POSIX"} or @code{"C"} lets
615 @code{string-collate-lessp} behave like @code{string-lessp}:
619 (sort '("11" "12" "1 1" "1 2" "1.1" "1.2")
620 (lambda (s1 s2) (string-collate-lessp s1 s2 "POSIX")))
621 @result{} ("1 1" "1 2" "1.1" "1.2" "11" "12")
625 If @var{ignore-case} is non-@code{nil}, characters are converted to lower-case
626 before comparing them.
628 To emulate Unicode-compliant collation on MS-Windows systems,
629 bind @code{w32-collate-ignore-punctuation} to a non-@code{nil} value, since
630 the codeset part of the locale cannot be @code{"UTF-8"} on MS-Windows.
632 If your system does not support a locale environment, this function
633 behaves like @code{string-lessp}.
636 @defun string-version-lessp string1 string2
637 This function compares strings lexicographically, except it treats
638 sequences of numerical characters as if they comprised a base-ten
639 number, and then compares the numbers. So @samp{foo2.png} is
640 ``smaller'' than @samp{foo12.png} according to this predicate, even if
641 @samp{12} is lexicographically ``smaller'' than @samp{2}.
644 @defun string-prefix-p string1 string2 &optional ignore-case
645 This function returns non-@code{nil} if @var{string1} is a prefix of
646 @var{string2}; i.e., if @var{string2} starts with @var{string1}. If
647 the optional argument @var{ignore-case} is non-@code{nil}, the
648 comparison ignores case differences.
651 @defun string-suffix-p suffix string &optional ignore-case
652 This function returns non-@code{nil} if @var{suffix} is a suffix of
653 @var{string}; i.e., if @var{string} ends with @var{suffix}. If the
654 optional argument @var{ignore-case} is non-@code{nil}, the comparison
655 ignores case differences.
658 @defun compare-strings string1 start1 end1 string2 start2 end2 &optional ignore-case
659 This function compares a specified part of @var{string1} with a
660 specified part of @var{string2}. The specified part of @var{string1}
661 runs from index @var{start1} (inclusive) up to index @var{end1}
662 (exclusive); @code{nil} for @var{start1} means the start of the
663 string, while @code{nil} for @var{end1} means the length of the
664 string. Likewise, the specified part of @var{string2} runs from index
665 @var{start2} up to index @var{end2}.
667 The strings are compared by the numeric values of their characters.
668 For instance, @var{str1} is considered less than @var{str2} if
669 its first differing character has a smaller numeric value. If
670 @var{ignore-case} is non-@code{nil}, characters are converted to
671 upper-case before comparing them. Unibyte strings are converted to
672 multibyte for comparison (@pxref{Text Representations}), so that a
673 unibyte string and its conversion to multibyte are always regarded as
676 If the specified portions of the two strings match, the value is
677 @code{t}. Otherwise, the value is an integer which indicates how many
678 leading characters agree, and which string is less. Its absolute
679 value is one plus the number of characters that agree at the beginning
680 of the two strings. The sign is negative if @var{string1} (or its
681 specified portion) is less.
684 @defun assoc-string key alist &optional case-fold
685 This function works like @code{assoc}, except that @var{key} must be a
686 string or symbol, and comparison is done using @code{compare-strings}.
687 Symbols are converted to strings before testing.
688 If @var{case-fold} is non-@code{nil}, @var{key} and the elements of
689 @var{alist} are converted to upper-case before comparison.
690 Unlike @code{assoc}, this function can also match elements of the alist
691 that are strings or symbols rather than conses. In particular, @var{alist} can
692 be a list of strings or symbols rather than an actual alist.
693 @xref{Association Lists}.
696 See also the function @code{compare-buffer-substrings} in
697 @ref{Comparing Text}, for a way to compare text in buffers. The
698 function @code{string-match}, which matches a regular expression
699 against a string, can be used for a kind of string comparison; see
702 @node String Conversion
703 @section Conversion of Characters and Strings
704 @cindex conversion of strings
706 This section describes functions for converting between characters,
707 strings and integers. @code{format} (@pxref{Formatting Strings}) and
708 @code{prin1-to-string} (@pxref{Output Functions}) can also convert
709 Lisp objects into strings. @code{read-from-string} (@pxref{Input
710 Functions}) can convert a string representation of a Lisp object
711 into an object. The functions @code{string-to-multibyte} and
712 @code{string-to-unibyte} convert the text representation of a string
713 (@pxref{Converting Representations}).
715 @xref{Documentation}, for functions that produce textual descriptions
716 of text characters and general input events
717 (@code{single-key-description} and @code{text-char-description}). These
718 are used primarily for making help messages.
720 @defun number-to-string number
721 @cindex integer to string
722 @cindex integer to decimal
723 This function returns a string consisting of the printed base-ten
724 representation of @var{number}. The returned value starts with a
725 minus sign if the argument is negative.
728 (number-to-string 256)
731 (number-to-string -23)
734 (number-to-string -23.5)
738 @cindex int-to-string
739 @code{int-to-string} is a semi-obsolete alias for this function.
741 See also the function @code{format} in @ref{Formatting Strings}.
744 @defun string-to-number string &optional base
745 @cindex string to number
746 This function returns the numeric value of the characters in
747 @var{string}. If @var{base} is non-@code{nil}, it must be an integer
748 between 2 and 16 (inclusive), and integers are converted in that base.
749 If @var{base} is @code{nil}, then base ten is used. Floating-point
750 conversion only works in base ten; we have not implemented other
751 radices for floating-point numbers, because that would be much more
752 work and does not seem useful. If @var{string} looks like an integer
753 but its value is too large to fit into a Lisp integer,
754 @code{string-to-number} returns a floating-point result.
756 The parsing skips spaces and tabs at the beginning of @var{string},
757 then reads as much of @var{string} as it can interpret as a number in
758 the given base. (On some systems it ignores other whitespace at the
759 beginning, not just spaces and tabs.) If @var{string} cannot be
760 interpreted as a number, this function returns 0.
763 (string-to-number "256")
765 (string-to-number "25 is a perfect square.")
767 (string-to-number "X256")
769 (string-to-number "-4.5")
771 (string-to-number "1e5")
775 @findex string-to-int
776 @code{string-to-int} is an obsolete alias for this function.
779 @defun char-to-string character
780 @cindex character to string
781 This function returns a new string containing one character,
782 @var{character}. This function is semi-obsolete because the function
783 @code{string} is more general. @xref{Creating Strings}.
786 @defun string-to-char string
787 This function returns the first character in @var{string}. This
788 mostly identical to @code{(aref string 0)}, except that it returns 0
789 if the string is empty. (The value is also 0 when the first character
790 of @var{string} is the null character, @acronym{ASCII} code 0.) This
791 function may be eliminated in the future if it does not seem useful
795 Here are some other functions that can convert to or from a string:
799 This function converts a vector or a list into a string.
800 @xref{Creating Strings}.
803 This function converts a string into a vector. @xref{Vector
807 This function converts a string into a list. @xref{Building Lists}.
810 This function converts a byte of character data into a unibyte string.
811 @xref{Converting Representations}.
814 @node Formatting Strings
815 @section Formatting Strings
816 @cindex formatting strings
817 @cindex strings, formatting them
819 @dfn{Formatting} means constructing a string by substituting
820 computed values at various places in a constant string. This constant
821 string controls how the other values are printed, as well as where
822 they appear; it is called a @dfn{format string}.
824 Formatting is often useful for computing messages to be displayed. In
825 fact, the functions @code{message} and @code{error} provide the same
826 formatting feature described here; they differ from @code{format-message} only
827 in how they use the result of formatting.
829 @defun format string &rest objects
830 This function returns a new string that is made by copying
831 @var{string} and then replacing any format specification
832 in the copy with encodings of the corresponding @var{objects}. The
833 arguments @var{objects} are the computed values to be formatted.
835 The characters in @var{string}, other than the format specifications,
836 are copied directly into the output, including their text properties,
837 if any. Any text properties of the format specifications are copied
838 to the produced string representations of the argument @var{objects}.
841 @defun format-message string &rest objects
842 @cindex curved quotes
844 This function acts like @code{format}, except it also converts any
845 grave accents (@t{`}) and apostrophes (@t{'}) in @var{string} as per the
846 value of @code{text-quoting-style}.
848 A format that quotes with grave accents and apostrophes @t{`like
849 this'} typically generates curved quotes @t{‘like this’}. In
850 contrast, a format that quotes with only apostrophes @t{'like this'}
851 typically generates two closing curved quotes @t{’like this’}, an
852 unusual style in English. @xref{Keys in Documentation}, for how the
853 @code{text-quoting-style} variable affects generated quotes.
856 @cindex @samp{%} in format
857 @cindex format specification
858 A format specification is a sequence of characters beginning with a
859 @samp{%}. Thus, if there is a @samp{%d} in @var{string}, the
860 @code{format} function replaces it with the printed representation of
861 one of the values to be formatted (one of the arguments @var{objects}).
866 (format "The value of fill-column is %d." fill-column)
867 @result{} "The value of fill-column is 72."
871 Since @code{format} interprets @samp{%} characters as format
872 specifications, you should @emph{never} pass an arbitrary string as
873 the first argument. This is particularly true when the string is
874 generated by some Lisp code. Unless the string is @emph{known} to
875 never include any @samp{%} characters, pass @code{"%s"}, described
876 below, as the first argument, and the string as the second, like this:
879 (format "%s" @var{arbitrary-string})
882 If @var{string} contains more than one format specification, the
883 format specifications correspond to successive values from
884 @var{objects}. Thus, the first format specification in @var{string}
885 uses the first such value, the second format specification uses the
886 second such value, and so on. Any extra format specifications (those
887 for which there are no corresponding values) cause an error. Any
888 extra values to be formatted are ignored.
890 Certain format specifications require values of particular types. If
891 you supply a value that doesn't fit the requirements, an error is
894 Here is a table of valid format specifications:
898 Replace the specification with the printed representation of the object,
899 made without quoting (that is, using @code{princ}, not
900 @code{prin1}---@pxref{Output Functions}). Thus, strings are represented
901 by their contents alone, with no @samp{"} characters, and symbols appear
902 without @samp{\} characters.
904 If the object is a string, its text properties are
905 copied into the output. The text properties of the @samp{%s} itself
906 are also copied, but those of the object take priority.
909 Replace the specification with the printed representation of the object,
910 made with quoting (that is, using @code{prin1}---@pxref{Output
911 Functions}). Thus, strings are enclosed in @samp{"} characters, and
912 @samp{\} characters appear where necessary before special characters.
915 @cindex integer to octal
916 Replace the specification with the base-eight representation of an
920 Replace the specification with the base-ten representation of an
925 @cindex integer to hexadecimal
926 Replace the specification with the base-sixteen representation of an
927 integer. @samp{%x} uses lower case and @samp{%X} uses upper case.
930 Replace the specification with the character which is the value given.
933 Replace the specification with the exponential notation for a
934 floating-point number.
937 Replace the specification with the decimal-point notation for a
938 floating-point number.
941 Replace the specification with notation for a floating-point number,
942 using either exponential notation or decimal-point notation, whichever
946 Replace the specification with a single @samp{%}. This format
947 specification is unusual in that it does not use a value. For example,
948 @code{(format "%% %d" 30)} returns @code{"% 30"}.
951 Any other format character results in an @samp{Invalid format
954 Here are several examples, which assume the typical
955 @code{text-quoting-style} settings:
959 (format "The octal value of %d is %o,
960 and the hex value is %x." 18 18 18)
961 @result{} "The octal value of 18 is 22,
962 and the hex value is 12."
965 "The name of this buffer is ‘%s’." (buffer-name))
966 @result{} "The name of this buffer is ‘strings.texi’."
969 "The buffer object prints as `%s'." (current-buffer))
970 @result{} "The buffer object prints as ‘strings.texi’."
976 A specification can have a @dfn{width}, which is a decimal number
977 between the @samp{%} and the specification character. If the printed
978 representation of the object contains fewer characters than this
979 width, @code{format} extends it with padding. The width specifier is
980 ignored for the @samp{%%} specification. Any padding introduced by
981 the width specifier normally consists of spaces inserted on the left:
984 (format "%5d is padded on the left with spaces" 123)
985 @result{} " 123 is padded on the left with spaces"
989 If the width is too small, @code{format} does not truncate the
990 object's printed representation. Thus, you can use a width to specify
991 a minimum spacing between columns with no risk of losing information.
992 In the following two examples, @samp{%7s} specifies a minimum width
993 of 7. In the first case, the string inserted in place of @samp{%7s}
994 has only 3 letters, and needs 4 blank spaces as padding. In the
995 second case, the string @code{"specification"} is 13 letters wide but
1000 (format "The word '%7s' has %d letters in it."
1001 "foo" (length "foo"))
1002 @result{} "The word ' foo' has 3 letters in it."
1003 (format "The word '%7s' has %d letters in it."
1004 "specification" (length "specification"))
1005 @result{} "The word 'specification' has 13 letters in it."
1009 @cindex flags in format specifications
1010 Immediately after the @samp{%} and before the optional width
1011 specifier, you can also put certain @dfn{flag characters}.
1013 The flag @samp{+} inserts a plus sign before a positive number, so
1014 that it always has a sign. A space character as flag inserts a space
1015 before a positive number. (Otherwise, positive numbers start with the
1016 first digit.) These flags are useful for ensuring that positive
1017 numbers and negative numbers use the same number of columns. They are
1018 ignored except for @samp{%d}, @samp{%e}, @samp{%f}, @samp{%g}, and if
1019 both flags are used, @samp{+} takes precedence.
1021 The flag @samp{#} specifies an alternate form which depends on
1022 the format in use. For @samp{%o}, it ensures that the result begins
1023 with a @samp{0}. For @samp{%x} and @samp{%X}, it prefixes the result
1024 with @samp{0x} or @samp{0X}. For @samp{%e}, @samp{%f}, and @samp{%g},
1025 the @samp{#} flag means include a decimal point even if the precision
1028 The flag @samp{0} ensures that the padding consists of @samp{0}
1029 characters instead of spaces. This flag is ignored for non-numerical
1030 specification characters like @samp{%s}, @samp{%S} and @samp{%c}.
1031 These specification characters accept the @samp{0} flag, but still pad
1034 The flag @samp{-} causes the padding inserted by the width
1035 specifier, if any, to be inserted on the right rather than the left.
1036 If both @samp{-} and @samp{0} are present, the @samp{0} flag is
1041 (format "%06d is padded on the left with zeros" 123)
1042 @result{} "000123 is padded on the left with zeros"
1044 (format "'%-6d' is padded on the right" 123)
1045 @result{} "'123 ' is padded on the right"
1047 (format "The word '%-7s' actually has %d letters in it."
1048 "foo" (length "foo"))
1049 @result{} "The word 'foo ' actually has 3 letters in it."
1053 @cindex precision in format specifications
1054 All the specification characters allow an optional @dfn{precision}
1055 before the character (after the width, if present). The precision is
1056 a decimal-point @samp{.} followed by a digit-string. For the
1057 floating-point specifications (@samp{%e}, @samp{%f}, @samp{%g}), the
1058 precision specifies how many decimal places to show; if zero, the
1059 decimal-point itself is also omitted. For @samp{%s} and @samp{%S},
1060 the precision truncates the string to the given width, so @samp{%.3s}
1061 shows only the first three characters of the representation for
1062 @var{object}. Precision has no effect for other specification
1065 @node Case Conversion
1066 @section Case Conversion in Lisp
1069 @cindex character case
1070 @cindex case conversion in Lisp
1072 The character case functions change the case of single characters or
1073 of the contents of strings. The functions normally convert only
1074 alphabetic characters (the letters @samp{A} through @samp{Z} and
1075 @samp{a} through @samp{z}, as well as non-@acronym{ASCII} letters); other
1076 characters are not altered. You can specify a different case
1077 conversion mapping by specifying a case table (@pxref{Case Tables}).
1079 These functions do not modify the strings that are passed to them as
1082 The examples below use the characters @samp{X} and @samp{x} which have
1083 @acronym{ASCII} codes 88 and 120 respectively.
1085 @defun downcase string-or-char
1086 This function converts @var{string-or-char}, which should be either a
1087 character or a string, to lower case.
1089 When @var{string-or-char} is a string, this function returns a new
1090 string in which each letter in the argument that is upper case is
1091 converted to lower case. When @var{string-or-char} is a character,
1092 this function returns the corresponding lower case character (an
1093 integer); if the original character is lower case, or is not a letter,
1094 the return value is equal to the original character.
1097 (downcase "The cat in the hat")
1098 @result{} "the cat in the hat"
1105 @defun upcase string-or-char
1106 This function converts @var{string-or-char}, which should be either a
1107 character or a string, to upper case.
1109 When @var{string-or-char} is a string, this function returns a new
1110 string in which each letter in the argument that is lower case is
1111 converted to upper case. When @var{string-or-char} is a character,
1112 this function returns the corresponding upper case character (an
1113 integer); if the original character is upper case, or is not a letter,
1114 the return value is equal to the original character.
1117 (upcase "The cat in the hat")
1118 @result{} "THE CAT IN THE HAT"
1125 @defun capitalize string-or-char
1126 @cindex capitalization
1127 This function capitalizes strings or characters. If
1128 @var{string-or-char} is a string, the function returns a new string
1129 whose contents are a copy of @var{string-or-char} in which each word
1130 has been capitalized. This means that the first character of each
1131 word is converted to upper case, and the rest are converted to lower
1134 The definition of a word is any sequence of consecutive characters that
1135 are assigned to the word constituent syntax class in the current syntax
1136 table (@pxref{Syntax Class Table}).
1138 When @var{string-or-char} is a character, this function does the same
1139 thing as @code{upcase}.
1143 (capitalize "The cat in the hat")
1144 @result{} "The Cat In The Hat"
1148 (capitalize "THE 77TH-HATTED CAT")
1149 @result{} "The 77th-Hatted Cat"
1159 @defun upcase-initials string-or-char
1160 If @var{string-or-char} is a string, this function capitalizes the
1161 initials of the words in @var{string-or-char}, without altering any
1162 letters other than the initials. It returns a new string whose
1163 contents are a copy of @var{string-or-char}, in which each word has
1164 had its initial letter converted to upper case.
1166 The definition of a word is any sequence of consecutive characters that
1167 are assigned to the word constituent syntax class in the current syntax
1168 table (@pxref{Syntax Class Table}).
1170 When the argument to @code{upcase-initials} is a character,
1171 @code{upcase-initials} has the same result as @code{upcase}.
1175 (upcase-initials "The CAT in the hAt")
1176 @result{} "The CAT In The HAt"
1181 @xref{Text Comparison}, for functions that compare strings; some of
1182 them ignore case differences, or can optionally ignore case differences.
1185 @section The Case Table
1187 You can customize case conversion by installing a special @dfn{case
1188 table}. A case table specifies the mapping between upper case and lower
1189 case letters. It affects both the case conversion functions for Lisp
1190 objects (see the previous section) and those that apply to text in the
1191 buffer (@pxref{Case Changes}). Each buffer has a case table; there is
1192 also a standard case table which is used to initialize the case table
1195 A case table is a char-table (@pxref{Char-Tables}) whose subtype is
1196 @code{case-table}. This char-table maps each character into the
1197 corresponding lower case character. It has three extra slots, which
1198 hold related tables:
1202 The upcase table maps each character into the corresponding upper
1205 The canonicalize table maps all of a set of case-related characters
1206 into a particular member of that set.
1208 The equivalences table maps each one of a set of case-related characters
1209 into the next character in that set.
1212 In simple cases, all you need to specify is the mapping to lower-case;
1213 the three related tables will be calculated automatically from that one.
1215 For some languages, upper and lower case letters are not in one-to-one
1216 correspondence. There may be two different lower case letters with the
1217 same upper case equivalent. In these cases, you need to specify the
1218 maps for both lower case and upper case.
1220 The extra table @var{canonicalize} maps each character to a canonical
1221 equivalent; any two characters that are related by case-conversion have
1222 the same canonical equivalent character. For example, since @samp{a}
1223 and @samp{A} are related by case-conversion, they should have the same
1224 canonical equivalent character (which should be either @samp{a} for both
1225 of them, or @samp{A} for both of them).
1227 The extra table @var{equivalences} is a map that cyclically permutes
1228 each equivalence class (of characters with the same canonical
1229 equivalent). (For ordinary @acronym{ASCII}, this would map @samp{a} into
1230 @samp{A} and @samp{A} into @samp{a}, and likewise for each set of
1231 equivalent characters.)
1233 When constructing a case table, you can provide @code{nil} for
1234 @var{canonicalize}; then Emacs fills in this slot from the lower case
1235 and upper case mappings. You can also provide @code{nil} for
1236 @var{equivalences}; then Emacs fills in this slot from
1237 @var{canonicalize}. In a case table that is actually in use, those
1238 components are non-@code{nil}. Do not try to specify
1239 @var{equivalences} without also specifying @var{canonicalize}.
1241 Here are the functions for working with case tables:
1243 @defun case-table-p object
1244 This predicate returns non-@code{nil} if @var{object} is a valid case
1248 @defun set-standard-case-table table
1249 This function makes @var{table} the standard case table, so that it will
1250 be used in any buffers created subsequently.
1253 @defun standard-case-table
1254 This returns the standard case table.
1257 @defun current-case-table
1258 This function returns the current buffer's case table.
1261 @defun set-case-table table
1262 This sets the current buffer's case table to @var{table}.
1265 @defmac with-case-table table body@dots{}
1266 The @code{with-case-table} macro saves the current case table, makes
1267 @var{table} the current case table, evaluates the @var{body} forms,
1268 and finally restores the case table. The return value is the value of
1269 the last form in @var{body}. The case table is restored even in case
1270 of an abnormal exit via @code{throw} or error (@pxref{Nonlocal
1274 Some language environments modify the case conversions of
1275 @acronym{ASCII} characters; for example, in the Turkish language
1276 environment, the @acronym{ASCII} capital I is downcased into
1277 a Turkish dotless i (@samp{ı}). This can interfere with code that requires
1278 ordinary @acronym{ASCII} case conversion, such as implementations of
1279 @acronym{ASCII}-based network protocols. In that case, use the
1280 @code{with-case-table} macro with the variable @var{ascii-case-table},
1281 which stores the unmodified case table for the @acronym{ASCII}
1284 @defvar ascii-case-table
1285 The case table for the @acronym{ASCII} character set. This should not be
1286 modified by any language environment settings.
1289 The following three functions are convenient subroutines for packages
1290 that define non-@acronym{ASCII} character sets. They modify the specified
1291 case table @var{case-table}; they also modify the standard syntax table.
1292 @xref{Syntax Tables}. Normally you would use these functions to change
1293 the standard case table.
1295 @defun set-case-syntax-pair uc lc case-table
1296 This function specifies a pair of corresponding letters, one upper case
1300 @defun set-case-syntax-delims l r case-table
1301 This function makes characters @var{l} and @var{r} a matching pair of
1302 case-invariant delimiters.
1305 @defun set-case-syntax char syntax case-table
1306 This function makes @var{char} case-invariant, with syntax
1310 @deffn Command describe-buffer-case-table
1311 This command displays a description of the contents of the current
1312 buffer's case table.