1 // © 2018 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
4 #include "unicode/utypes.h"
6 #if !UCONFIG_NO_FORMATTING
7 #ifndef __UNUMBERFORMATTER_H__
8 #define __UNUMBERFORMATTER_H__
10 #include "unicode/parseerr.h"
11 #include "unicode/ufieldpositer.h"
12 #include "unicode/umisc.h"
13 #include "unicode/uformattedvalue.h"
18 * \brief C-compatible API for localized number formatting; not recommended for C++.
20 * This is the C-compatible version of the NumberFormatter API introduced in ICU 60. C++ users should
21 * include unicode/numberformatter.h and use the proper C++ APIs.
23 * The C API accepts a number skeleton string for specifying the settings for formatting, which covers a
24 * very large subset of all possible number formatting features. For more information on number skeleton
25 * strings, see unicode/numberformatter.h.
27 * When using UNumberFormatter, which is treated as immutable, the results are exported to a mutable
28 * UFormattedNumber object, which you subsequently use for populating your string buffer or iterating over
34 * UErrorCode ec = U_ZERO_ERROR;
35 * UNumberFormatter* uformatter = unumf_openForSkeletonAndLocale(u"precision-integer", -1, "en", &ec);
36 * UFormattedNumber* uresult = unumf_openResult(&ec);
37 * if (U_FAILURE(ec)) { return; }
40 * unumf_formatDouble(uformatter, 5142.3, uresult, &ec);
41 * if (U_FAILURE(ec)) { return; }
43 * // Export the string to a malloc'd buffer:
44 * int32_t len = unumf_resultToString(uresult, NULL, 0, &ec);
45 * // at this point, ec == U_BUFFER_OVERFLOW_ERROR
47 * UChar* buffer = (UChar*) malloc((len+1)*sizeof(UChar));
48 * unumf_resultToString(uresult, buffer, len+1, &ec);
49 * if (U_FAILURE(ec)) { return; }
50 * // buffer should equal "5,142"
53 * unumf_close(uformatter);
54 * unumf_closeResult(uresult);
58 * If you are a C++ user linking against the C libraries, you can use the LocalPointer versions of these
59 * APIs. The following example uses LocalPointer with the decimal number and field position APIs:
63 * LocalUNumberFormatterPointer uformatter(unumf_openForSkeletonAndLocale(u"percent", -1, "en", &ec));
64 * LocalUFormattedNumberPointer uresult(unumf_openResult(&ec));
65 * if (U_FAILURE(ec)) { return; }
67 * // Format a decimal number:
68 * unumf_formatDecimal(uformatter.getAlias(), "9.87E-3", -1, uresult.getAlias(), &ec);
69 * if (U_FAILURE(ec)) { return; }
71 * // Get the location of the percent sign:
72 * UFieldPosition ufpos = {UNUM_PERCENT_FIELD, 0, 0};
73 * unumf_resultNextFieldPosition(uresult.getAlias(), &ufpos, &ec);
74 * // ufpos should contain beginIndex=7 and endIndex=8 since the string is "0.00987%"
76 * // No need to do any cleanup since we are using LocalPointer.
81 * An enum declaring how to render units, including currencies. Example outputs when formatting 123 USD and 123
82 * meters in <em>en-CA</em>:
86 * <li>NARROW*: "$123.00" and "123 m"
87 * <li>SHORT: "US$ 123.00" and "123 m"
88 * <li>FULL_NAME: "123.00 US dollars" and "123 meters"
89 * <li>ISO_CODE: "USD 123.00" and undefined behavior
90 * <li>HIDDEN: "123.00" and "123"
94 * This enum is similar to {@link UMeasureFormatWidth}.
98 typedef enum UNumberUnitWidth
{
100 * Print an abbreviated version of the unit name. Similar to SHORT, but always use the shortest available
101 * abbreviation or symbol. This option can be used when the context hints at the identity of the unit. For more
102 * information on the difference between NARROW and SHORT, see SHORT.
105 * In CLDR, this option corresponds to the "Narrow" format for measure units and the "¤¤¤¤¤" placeholder for
110 UNUM_UNIT_WIDTH_NARROW
,
113 * Print an abbreviated version of the unit name. Similar to NARROW, but use a slightly wider abbreviation or
114 * symbol when there may be ambiguity. This is the default behavior.
117 * For example, in <em>es-US</em>, the SHORT form for Fahrenheit is "{0} °F", but the NARROW form is "{0}°",
118 * since Fahrenheit is the customary unit for temperature in that locale.
121 * In CLDR, this option corresponds to the "Short" format for measure units and the "¤" placeholder for
126 UNUM_UNIT_WIDTH_SHORT
,
129 * Print the full name of the unit, without any abbreviations.
132 * In CLDR, this option corresponds to the default format for measure units and the "¤¤¤" placeholder for
137 UNUM_UNIT_WIDTH_FULL_NAME
,
140 * Use the three-digit ISO XXX code in place of the symbol for displaying currencies. The behavior of this
141 * option is currently undefined for use with measure units.
144 * In CLDR, this option corresponds to the "¤¤" placeholder for currencies.
148 UNUM_UNIT_WIDTH_ISO_CODE
,
151 * Format the number according to the specified unit, but do not display the unit. For currencies, apply
152 * monetary symbols and formats as with SHORT, but omit the currency symbol. For measure units, the behavior is
153 * equivalent to not specifying the unit at all.
157 UNUM_UNIT_WIDTH_HIDDEN
,
160 * One more than the highest UNumberUnitWidth value.
162 * @internal ICU 60: The numeric value may change over time; see ICU ticket #12420.
164 UNUM_UNIT_WIDTH_COUNT
168 * An enum declaring the strategy for when and how to display grouping separators (i.e., the
169 * separator, often a comma or period, after every 2-3 powers of ten). The choices are several
170 * pre-built strategies for different use cases that employ locale data whenever possible. Example
171 * outputs for 1234 and 1234567 in <em>en-IN</em>:
174 * <li>OFF: 1234 and 12345
175 * <li>MIN2: 1234 and 12,34,567
176 * <li>AUTO: 1,234 and 12,34,567
177 * <li>ON_ALIGNED: 1,234 and 12,34,567
178 * <li>THOUSANDS: 1,234 and 1,234,567
182 * The default is AUTO, which displays grouping separators unless the locale data says that grouping
183 * is not customary. To force grouping for all numbers greater than 1000 consistently across locales,
184 * use ON_ALIGNED. On the other hand, to display grouping less frequently than the default, use MIN2
185 * or OFF. See the docs of each option for details.
188 * Note: This enum specifies the strategy for grouping sizes. To set which character to use as the
189 * grouping separator, use the "symbols" setter.
193 typedef enum UNumberGroupingStrategy
{
195 * Do not display grouping separators in any locale.
202 * Display grouping using locale defaults, except do not show grouping on values smaller than
203 * 10000 (such that there is a <em>minimum of two digits</em> before the first separator).
206 * Note that locales may restrict grouping separators to be displayed only on 1 million or
207 * greater (for example, ee and hu) or disable grouping altogether (for example, bg currency).
210 * Locale data is used to determine whether to separate larger numbers into groups of 2
211 * (customary in South Asia) or groups of 3 (customary in Europe and the Americas).
218 * Display grouping using the default strategy for all locales. This is the default behavior.
221 * Note that locales may restrict grouping separators to be displayed only on 1 million or
222 * greater (for example, ee and hu) or disable grouping altogether (for example, bg currency).
225 * Locale data is used to determine whether to separate larger numbers into groups of 2
226 * (customary in South Asia) or groups of 3 (customary in Europe and the Americas).
233 * Always display the grouping separator on values of at least 1000.
236 * This option ignores the locale data that restricts or disables grouping, described in MIN2 and
237 * AUTO. This option may be useful to normalize the alignment of numbers, such as in a
241 * Locale data is used to determine whether to separate larger numbers into groups of 2
242 * (customary in South Asia) or groups of 3 (customary in Europe and the Americas).
246 UNUM_GROUPING_ON_ALIGNED
,
249 * Use the Western defaults: groups of 3 and enabled for all numbers 1000 or greater. Do not use
250 * locale data for determining the grouping strategy.
254 UNUM_GROUPING_THOUSANDS
256 #ifndef U_HIDE_INTERNAL_API
259 * One more than the highest UNumberGroupingStrategy value.
261 * @internal ICU 62: The numeric value may change over time; see ICU ticket #12420.
264 #endif /* U_HIDE_INTERNAL_API */
266 } UNumberGroupingStrategy
;
269 * An enum declaring how to denote positive and negative numbers. Example outputs when formatting
270 * 123, 0, and -123 in <em>en-US</em>:
273 * <li>AUTO: "123", "0", and "-123"
274 * <li>ALWAYS: "+123", "+0", and "-123"
275 * <li>NEVER: "123", "0", and "123"
276 * <li>ACCOUNTING: "$123", "$0", and "($123)"
277 * <li>ACCOUNTING_ALWAYS: "+$123", "+$0", and "($123)"
278 * <li>EXCEPT_ZERO: "+123", "0", and "-123"
279 * <li>ACCOUNTING_EXCEPT_ZERO: "+$123", "$0", and "($123)"
283 * The exact format, including the position and the code point of the sign, differ by locale.
287 typedef enum UNumberSignDisplay
{
289 * Show the minus sign on negative numbers, and do not show the sign on positive numbers. This is the default
297 * Show the minus sign on negative numbers and the plus sign on positive numbers, including zero.
298 * To hide the sign on zero, see {@link UNUM_SIGN_EXCEPT_ZERO}.
305 * Do not show the sign on positive or negative numbers.
312 * Use the locale-dependent accounting format on negative numbers, and do not show the sign on positive numbers.
315 * The accounting format is defined in CLDR and varies by locale; in many Western locales, the format is a pair
316 * of parentheses around the number.
319 * Note: Since CLDR defines the accounting format in the monetary context only, this option falls back to the
320 * AUTO sign display strategy when formatting without a currency unit. This limitation may be lifted in the
325 UNUM_SIGN_ACCOUNTING
,
328 * Use the locale-dependent accounting format on negative numbers, and show the plus sign on
329 * positive numbers, including zero. For more information on the accounting format, see the
330 * ACCOUNTING sign display strategy. To hide the sign on zero, see
331 * {@link UNUM_SIGN_ACCOUNTING_EXCEPT_ZERO}.
335 UNUM_SIGN_ACCOUNTING_ALWAYS
,
338 * Show the minus sign on negative numbers and the plus sign on positive numbers. Do not show a
339 * sign on zero or NaN, unless the sign bit is set (-0.0 gets a sign).
343 UNUM_SIGN_EXCEPT_ZERO
,
346 * Use the locale-dependent accounting format on negative numbers, and show the plus sign on
347 * positive numbers. Do not show a sign on zero or NaN, unless the sign bit is set (-0.0 gets a
348 * sign). For more information on the accounting format, see the ACCOUNTING sign display
353 UNUM_SIGN_ACCOUNTING_EXCEPT_ZERO
,
356 * One more than the highest UNumberSignDisplay value.
358 * @internal ICU 60: The numeric value may change over time; see ICU ticket #12420.
361 } UNumberSignDisplay
;
364 * An enum declaring how to render the decimal separator.
368 * <li>UNUM_DECIMAL_SEPARATOR_AUTO: "1", "1.1"
369 * <li>UNUM_DECIMAL_SEPARATOR_ALWAYS: "1.", "1.1"
374 typedef enum UNumberDecimalSeparatorDisplay
{
376 * Show the decimal separator when there are one or more digits to display after the separator, and do not show
377 * it otherwise. This is the default behavior.
381 UNUM_DECIMAL_SEPARATOR_AUTO
,
384 * Always show the decimal separator, even if there are no digits to display after the separator.
388 UNUM_DECIMAL_SEPARATOR_ALWAYS
,
391 * One more than the highest UNumberDecimalSeparatorDisplay value.
393 * @internal ICU 60: The numeric value may change over time; see ICU ticket #12420.
395 UNUM_DECIMAL_SEPARATOR_COUNT
396 } UNumberDecimalSeparatorDisplay
;
398 struct UNumberFormatter
;
400 * C-compatible version of icu::number::LocalizedNumberFormatter.
402 * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
406 typedef struct UNumberFormatter UNumberFormatter
;
408 struct UFormattedNumber
;
410 * C-compatible version of icu::number::FormattedNumber.
412 * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
416 typedef struct UFormattedNumber UFormattedNumber
;
420 * Creates a new UNumberFormatter for the given skeleton string and locale. This is currently the only
421 * method for creating a new UNumberFormatter.
423 * Objects of type UNumberFormatter returned by this method are threadsafe.
425 * For more details on skeleton strings, see the documentation in numberformatter.h. For more details on
426 * the usage of this API, see the documentation at the top of unumberformatter.h.
428 * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
430 * @param skeleton The skeleton string, like u"percent precision-integer"
431 * @param skeletonLen The number of UChars in the skeleton string, or -1 it it is NUL-terminated.
432 * @param locale The NUL-terminated locale ID.
433 * @param ec Set if an error occurs.
436 U_STABLE UNumberFormatter
* U_EXPORT2
437 unumf_openForSkeletonAndLocale(const UChar
* skeleton
, int32_t skeletonLen
, const char* locale
,
441 #ifndef U_HIDE_DRAFT_API
443 * Like unumf_openForSkeletonAndLocale, but accepts a UParseError, which will be populated with the
444 * location of a skeleton syntax error if such a syntax error exists.
446 * @param skeleton The skeleton string, like u"percent precision-integer"
447 * @param skeletonLen The number of UChars in the skeleton string, or -1 it it is NUL-terminated.
448 * @param locale The NUL-terminated locale ID.
449 * @param perror A parse error struct populated if an error occurs when parsing. Can be NULL.
450 * If no error occurs, perror->offset will be set to -1.
451 * @param ec Set if an error occurs.
454 U_DRAFT UNumberFormatter
* U_EXPORT2
455 unumf_openForSkeletonAndLocaleWithError(
456 const UChar
* skeleton
, int32_t skeletonLen
, const char* locale
, UParseError
* perror
, UErrorCode
* ec
);
457 #endif // U_HIDE_DRAFT_API
461 * Creates an object to hold the result of a UNumberFormatter
462 * operation. The object can be used repeatedly; it is cleared whenever
463 * passed to a format function.
465 * @param ec Set if an error occurs.
468 U_STABLE UFormattedNumber
* U_EXPORT2
469 unumf_openResult(UErrorCode
* ec
);
473 * Uses a UNumberFormatter to format an integer to a UFormattedNumber. A string, field position, and other
474 * information can be retrieved from the UFormattedNumber.
476 * The UNumberFormatter can be shared between threads. Each thread should have its own local
477 * UFormattedNumber, however, for storing the result of the formatting operation.
479 * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
481 * @param uformatter A formatter object created by unumf_openForSkeletonAndLocale or similar.
482 * @param value The number to be formatted.
483 * @param uresult The object that will be mutated to store the result; see unumf_openResult.
484 * @param ec Set if an error occurs.
487 U_STABLE
void U_EXPORT2
488 unumf_formatInt(const UNumberFormatter
* uformatter
, int64_t value
, UFormattedNumber
* uresult
,
493 * Uses a UNumberFormatter to format a double to a UFormattedNumber. A string, field position, and other
494 * information can be retrieved from the UFormattedNumber.
496 * The UNumberFormatter can be shared between threads. Each thread should have its own local
497 * UFormattedNumber, however, for storing the result of the formatting operation.
499 * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
501 * @param uformatter A formatter object created by unumf_openForSkeletonAndLocale or similar.
502 * @param value The number to be formatted.
503 * @param uresult The object that will be mutated to store the result; see unumf_openResult.
504 * @param ec Set if an error occurs.
507 U_STABLE
void U_EXPORT2
508 unumf_formatDouble(const UNumberFormatter
* uformatter
, double value
, UFormattedNumber
* uresult
,
513 * Uses a UNumberFormatter to format a decimal number to a UFormattedNumber. A string, field position, and
514 * other information can be retrieved from the UFormattedNumber.
516 * The UNumberFormatter can be shared between threads. Each thread should have its own local
517 * UFormattedNumber, however, for storing the result of the formatting operation.
519 * The syntax of the unformatted number is a "numeric string" as defined in the Decimal Arithmetic
520 * Specification, available at http://speleotrove.com/decimal
522 * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
524 * @param uformatter A formatter object created by unumf_openForSkeletonAndLocale or similar.
525 * @param value The numeric string to be formatted.
526 * @param valueLen The length of the numeric string, or -1 if it is NUL-terminated.
527 * @param uresult The object that will be mutated to store the result; see unumf_openResult.
528 * @param ec Set if an error occurs.
531 U_STABLE
void U_EXPORT2
532 unumf_formatDecimal(const UNumberFormatter
* uformatter
, const char* value
, int32_t valueLen
,
533 UFormattedNumber
* uresult
, UErrorCode
* ec
);
535 #ifndef U_HIDE_DRAFT_API
537 * Returns a representation of a UFormattedNumber as a UFormattedValue,
538 * which can be subsequently passed to any API requiring that type.
540 * The returned object is owned by the UFormattedNumber and is valid
541 * only as long as the UFormattedNumber is present and unchanged in memory.
543 * You can think of this method as a cast between types.
545 * @param uresult The object containing the formatted string.
546 * @param ec Set if an error occurs.
547 * @return A UFormattedValue owned by the input object.
550 U_DRAFT
const UFormattedValue
* U_EXPORT2
551 unumf_resultAsValue(const UFormattedNumber
* uresult
, UErrorCode
* ec
);
552 #endif /* U_HIDE_DRAFT_API */
556 * Extracts the result number string out of a UFormattedNumber to a UChar buffer if possible.
557 * If bufferCapacity is greater than the required length, a terminating NUL is written.
558 * If bufferCapacity is less than the required length, an error code is set.
560 * Also see ufmtval_getString, which returns a NUL-terminated string:
563 * const UChar* str = ufmtval_getString(unumf_resultAsValue(uresult, &ec), &len, &ec);
565 * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
567 * @param uresult The object containing the formatted number.
568 * @param buffer Where to save the string output.
569 * @param bufferCapacity The number of UChars available in the buffer.
570 * @param ec Set if an error occurs.
571 * @return The required length.
574 U_STABLE
int32_t U_EXPORT2
575 unumf_resultToString(const UFormattedNumber
* uresult
, UChar
* buffer
, int32_t bufferCapacity
,
580 * Determines the start and end indices of the next occurrence of the given <em>field</em> in the
581 * output string. This allows you to determine the locations of, for example, the integer part,
582 * fraction part, or symbols.
584 * This is a simpler but less powerful alternative to {@link ufmtval_nextPosition}.
586 * If a field occurs just once, calling this method will find that occurrence and return it. If a
587 * field occurs multiple times, this method may be called repeatedly with the following pattern:
590 * UFieldPosition ufpos = {UNUM_GROUPING_SEPARATOR_FIELD, 0, 0};
591 * while (unumf_resultNextFieldPosition(uresult, ufpos, &ec)) {
592 * // do something with ufpos.
596 * This method is useful if you know which field to query. If you want all available field position
597 * information, use unumf_resultGetAllFieldPositions().
599 * NOTE: All fields of the UFieldPosition must be initialized before calling this method.
601 * @param uresult The object containing the formatted number.
603 * Input+output variable. On input, the "field" property determines which field to look up,
604 * and the "endIndex" property determines where to begin the search. On output, the
605 * "beginIndex" field is set to the beginning of the first occurrence of the field after the
606 * input "endIndex", and "endIndex" is set to the end of that occurrence of the field
607 * (exclusive index). If a field position is not found, the FieldPosition is not changed and
608 * the method returns FALSE.
609 * @param ec Set if an error occurs.
612 U_STABLE UBool U_EXPORT2
613 unumf_resultNextFieldPosition(const UFormattedNumber
* uresult
, UFieldPosition
* ufpos
, UErrorCode
* ec
);
617 * Populates the given iterator with all fields in the formatted output string. This allows you to
618 * determine the locations of the integer part, fraction part, and sign.
620 * This is an alternative to the more powerful {@link ufmtval_nextPosition} API.
622 * If you need information on only one field, use {@link ufmtval_nextPosition} or
623 * {@link unumf_resultNextFieldPosition}.
625 * @param uresult The object containing the formatted number.
627 * A pointer to a UFieldPositionIterator created by {@link #ufieldpositer_open}. Iteration
628 * information already present in the UFieldPositionIterator is deleted, and the iterator is reset
629 * to apply to the fields in the formatted string created by this function call. The field values
630 * and indexes returned by {@link #ufieldpositer_next} represent fields denoted by
631 * the UNumberFormatFields enum. Fields are not returned in a guaranteed order. Fields cannot
632 * overlap, but they may nest. For example, 1234 could format as "1,234" which might consist of a
633 * grouping separator field for ',' and an integer field encompassing the entire string.
634 * @param ec Set if an error occurs.
637 U_STABLE
void U_EXPORT2
638 unumf_resultGetAllFieldPositions(const UFormattedNumber
* uresult
, UFieldPositionIterator
* ufpositer
,
642 // TODO(ICU-20775): Propose this as API.
643 // NOTE: This is not currently implemented.
644 // U_DRAFT int32_t U_EXPORT2
645 // unumf_resultToDecimalNumber(const UFormattedNumber* uresult, char* buffer, int32_t bufferCapacity,
650 * Releases the UNumberFormatter created by unumf_openForSkeletonAndLocale().
652 * @param uformatter An object created by unumf_openForSkeletonAndLocale().
655 U_STABLE
void U_EXPORT2
656 unumf_close(UNumberFormatter
* uformatter
);
660 * Releases the UFormattedNumber created by unumf_openResult().
662 * @param uresult An object created by unumf_openResult().
665 U_STABLE
void U_EXPORT2
666 unumf_closeResult(UFormattedNumber
* uresult
);
669 #if U_SHOW_CPLUSPLUS_API
673 * \class LocalUNumberFormatterPointer
674 * "Smart pointer" class; closes a UNumberFormatter via unumf_close().
675 * For most methods see the LocalPointerBase base class.
679 * LocalUNumberFormatterPointer uformatter(unumf_openForSkeletonAndLocale(...));
680 * // no need to explicitly call unumf_close()
683 * @see LocalPointerBase
687 U_DEFINE_LOCAL_OPEN_POINTER(LocalUNumberFormatterPointer
, UNumberFormatter
, unumf_close
);
690 * \class LocalUFormattedNumberPointer
691 * "Smart pointer" class; closes a UFormattedNumber via unumf_closeResult().
692 * For most methods see the LocalPointerBase base class.
696 * LocalUFormattedNumberPointer uformatter(unumf_openResult(...));
697 * // no need to explicitly call unumf_closeResult()
700 * @see LocalPointerBase
704 U_DEFINE_LOCAL_OPEN_POINTER(LocalUFormattedNumberPointer
, UFormattedNumber
, unumf_closeResult
);
707 #endif // U_SHOW_CPLUSPLUS_API
709 #endif //__UNUMBERFORMATTER_H__
710 #endif /* #if !UCONFIG_NO_FORMATTING */