xref: /aosp_15_r20/external/icu/libandroidicu/include/unicode/unumberformatter.h (revision 0e209d3975ff4a8c132096b14b0e9364a753506e)
1 // © 2018 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
3 
4 #ifndef __UNUMBERFORMATTER_H__
5 #define __UNUMBERFORMATTER_H__
6 
7 #include "unicode/utypes.h"
8 
9 #if !UCONFIG_NO_FORMATTING
10 
11 #include "unicode/parseerr.h"
12 #include "unicode/unumberoptions.h"
13 #include "unicode/uformattednumber.h"
14 
15 
16 /**
17  * \file
18  * \brief C API: Localized number formatting; not recommended for C++.
19  *
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.
22  *
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.
26  *
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
29  * the fields.
30  *
31  * Example code:
32  * <pre>
33  * // Setup:
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; }
38  *
39  * // Format a double:
40  * unumf_formatDouble(uformatter, 5142.3, uresult, &ec);
41  * if (U_FAILURE(ec)) { return; }
42  *
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
46  * ec = U_ZERO_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"
51  *
52  * // Cleanup:
53  * unumf_close(uformatter);
54  * unumf_closeResult(uresult);
55  * free(buffer);
56  * </pre>
57  *
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:
60  *
61  * <pre>
62  * // Setup:
63  * LocalUNumberFormatterPointer uformatter(unumf_openForSkeletonAndLocale(u"percent", -1, "en", &ec));
64  * LocalUFormattedNumberPointer uresult(unumf_openResult(&ec));
65  * if (U_FAILURE(ec)) { return; }
66  *
67  * // Format a decimal number:
68  * unumf_formatDecimal(uformatter.getAlias(), "9.87E-3", -1, uresult.getAlias(), &ec);
69  * if (U_FAILURE(ec)) { return; }
70  *
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%"
75  *
76  * // No need to do any cleanup since we are using LocalPointer.
77  * </pre>
78  */
79 
80 /**
81  * An enum declaring how to resolve conflicts between maximum fraction digits and maximum
82  * significant digits.
83  *
84  * There are two modes, RELAXED and STRICT:
85  *
86  * - RELAXED: Relax one of the two constraints (fraction digits or significant digits) in order
87  *   to round the number to a higher level of precision.
88  * - STRICT: Enforce both constraints, resulting in the number being rounded to a lower
89  *   level of precision.
90  *
91  * The default settings for compact notation rounding are Max-Fraction = 0 (round to the nearest
92  * integer), Max-Significant = 2 (round to 2 significant digits), and priority RELAXED (choose
93  * the constraint that results in more digits being displayed).
94  *
95  * Conflicting *minimum* fraction and significant digits are always resolved in the direction that
96  * results in more trailing zeros.
97  *
98  * Example 1: Consider the number 3.141, with various different settings:
99  *
100  * - Max-Fraction = 1: "3.1"
101  * - Max-Significant = 3: "3.14"
102  *
103  * The rounding priority determines how to resolve the conflict when both Max-Fraction and
104  * Max-Significant are set. With RELAXED, the less-strict setting (the one that causes more digits
105  * to be displayed) will be used; Max-Significant wins. With STRICT, the more-strict setting (the
106  * one that causes fewer digits to be displayed) will be used; Max-Fraction wins.
107  *
108  * Example 2: Consider the number 8317, with various different settings:
109  *
110  * - Max-Fraction = 1: "8317"
111  * - Max-Significant = 3: "8320"
112  *
113  * Here, RELAXED favors Max-Fraction and STRICT favors Max-Significant. Note that this larger
114  * number caused the two modes to favor the opposite result.
115  *
116  * @stable ICU 69
117  */
118 typedef enum UNumberRoundingPriority {
119     /**
120      * Favor greater precision by relaxing one of the rounding constraints.
121      *
122      * @stable ICU 69
123      */
124     UNUM_ROUNDING_PRIORITY_RELAXED,
125 
126     /**
127      * Favor adherence to all rounding constraints by producing lower precision.
128      *
129      * @stable ICU 69
130      */
131     UNUM_ROUNDING_PRIORITY_STRICT,
132 } UNumberRoundingPriority;
133 
134 /**
135  * An enum declaring how to render units, including currencies. Example outputs when formatting 123 USD and 123
136  * meters in <em>en-CA</em>:
137  *
138  * <p>
139  * <ul>
140  * <li>NARROW*: "$123.00" and "123 m"
141  * <li>SHORT: "US$ 123.00" and "123 m"
142  * <li>FULL_NAME: "123.00 US dollars" and "123 meters"
143  * <li>ISO_CODE: "USD 123.00" and undefined behavior
144  * <li>HIDDEN: "123.00" and "123"
145  * </ul>
146  *
147  * <p>
148  * This enum is similar to {@link UMeasureFormatWidth}.
149  *
150  * @stable ICU 60
151  */
152 typedef enum UNumberUnitWidth {
153     /**
154      * Print an abbreviated version of the unit name. Similar to SHORT, but always use the shortest available
155      * abbreviation or symbol. This option can be used when the context hints at the identity of the unit. For more
156      * information on the difference between NARROW and SHORT, see SHORT.
157      *
158      * <p>
159      * In CLDR, this option corresponds to the "Narrow" format for measure units and the "¤¤¤¤¤" placeholder for
160      * currencies.
161      *
162      * @stable ICU 60
163      */
164             UNUM_UNIT_WIDTH_NARROW = 0,
165 
166     /**
167      * Print an abbreviated version of the unit name. Similar to NARROW, but use a slightly wider abbreviation or
168      * symbol when there may be ambiguity. This is the default behavior.
169      *
170      * <p>
171      * For example, in <em>es-US</em>, the SHORT form for Fahrenheit is "{0} °F", but the NARROW form is "{0}°",
172      * since Fahrenheit is the customary unit for temperature in that locale.
173      *
174      * <p>
175      * In CLDR, this option corresponds to the "Short" format for measure units and the "¤" placeholder for
176      * currencies.
177      *
178      * @stable ICU 60
179      */
180             UNUM_UNIT_WIDTH_SHORT = 1,
181 
182     /**
183      * Print the full name of the unit, without any abbreviations.
184      *
185      * <p>
186      * In CLDR, this option corresponds to the default format for measure units and the "¤¤¤" placeholder for
187      * currencies.
188      *
189      * @stable ICU 60
190      */
191             UNUM_UNIT_WIDTH_FULL_NAME = 2,
192 
193     /**
194      * Use the three-digit ISO XXX code in place of the symbol for displaying currencies. The behavior of this
195      * option is currently undefined for use with measure units.
196      *
197      * <p>
198      * In CLDR, this option corresponds to the "¤¤" placeholder for currencies.
199      *
200      * @stable ICU 60
201      */
202             UNUM_UNIT_WIDTH_ISO_CODE = 3,
203 
204     /**
205      * Use the formal variant of the currency symbol; for example, "NT$" for the New Taiwan
206      * dollar in zh-TW.
207      *
208      * <p>
209      * Behavior of this option with non-currency units is not defined at this time.
210      *
211      * @stable ICU 68
212      */
213             UNUM_UNIT_WIDTH_FORMAL = 4,
214 
215     /**
216      * Use the alternate variant of the currency symbol; for example, "TL" for the Turkish
217      * lira (TRY).
218      *
219      * <p>
220      * Behavior of this option with non-currency units is not defined at this time.
221      *
222      * @stable ICU 68
223      */
224             UNUM_UNIT_WIDTH_VARIANT = 5,
225 
226     /**
227      * Format the number according to the specified unit, but do not display the unit. For currencies, apply
228      * monetary symbols and formats as with SHORT, but omit the currency symbol. For measure units, the behavior is
229      * equivalent to not specifying the unit at all.
230      *
231      * @stable ICU 60
232      */
233             UNUM_UNIT_WIDTH_HIDDEN = 6,
234 
235     // Do not conditionalize the following with #ifndef U_HIDE_INTERNAL_API,
236     // needed for unconditionalized struct MacroProps
237     /**
238      * One more than the highest UNumberUnitWidth value.
239      *
240      * @internal ICU 60: The numeric value may change over time; see ICU ticket #12420.
241      */
242             UNUM_UNIT_WIDTH_COUNT = 7
243 } UNumberUnitWidth;
244 
245 /**
246  * An enum declaring how to denote positive and negative numbers. Example outputs when formatting
247  * 123, 0, and -123 in <em>en-US</em>:
248  *
249  * <ul>
250  * <li>AUTO: "123", "0", and "-123"
251  * <li>ALWAYS: "+123", "+0", and "-123"
252  * <li>NEVER: "123", "0", and "123"
253  * <li>ACCOUNTING: "$123", "$0", and "($123)"
254  * <li>ACCOUNTING_ALWAYS: "+$123", "+$0", and "($123)"
255  * <li>EXCEPT_ZERO: "+123", "0", and "-123"
256  * <li>ACCOUNTING_EXCEPT_ZERO: "+$123", "$0", and "($123)"
257  * </ul>
258  *
259  * <p>
260  * The exact format, including the position and the code point of the sign, differ by locale.
261  *
262  * @stable ICU 60
263  */
264 typedef enum UNumberSignDisplay {
265     /**
266      * Show the minus sign on negative numbers, and do not show the sign on positive numbers. This is the default
267      * behavior.
268      *
269      * If using this option, a sign will be displayed on negative zero, including negative numbers
270      * that round to zero. To hide the sign on negative zero, use the NEGATIVE option.
271      *
272      * @stable ICU 60
273      */
274     UNUM_SIGN_AUTO,
275 
276     /**
277      * Show the minus sign on negative numbers and the plus sign on positive numbers, including zero.
278      * To hide the sign on zero, see {@link UNUM_SIGN_EXCEPT_ZERO}.
279      *
280      * @stable ICU 60
281      */
282     UNUM_SIGN_ALWAYS,
283 
284     /**
285      * Do not show the sign on positive or negative numbers.
286      *
287      * @stable ICU 60
288      */
289     UNUM_SIGN_NEVER,
290 
291     /**
292      * Use the locale-dependent accounting format on negative numbers, and do not show the sign on positive numbers.
293      *
294      * <p>
295      * The accounting format is defined in CLDR and varies by locale; in many Western locales, the format is a pair
296      * of parentheses around the number.
297      *
298      * <p>
299      * Note: Since CLDR defines the accounting format in the monetary context only, this option falls back to the
300      * AUTO sign display strategy when formatting without a currency unit. This limitation may be lifted in the
301      * future.
302      *
303      * @stable ICU 60
304      */
305     UNUM_SIGN_ACCOUNTING,
306 
307     /**
308      * Use the locale-dependent accounting format on negative numbers, and show the plus sign on
309      * positive numbers, including zero. For more information on the accounting format, see the
310      * ACCOUNTING sign display strategy. To hide the sign on zero, see
311      * {@link UNUM_SIGN_ACCOUNTING_EXCEPT_ZERO}.
312      *
313      * @stable ICU 60
314      */
315     UNUM_SIGN_ACCOUNTING_ALWAYS,
316 
317     /**
318      * Show the minus sign on negative numbers and the plus sign on positive numbers. Do not show a
319      * sign on zero, numbers that round to zero, or NaN.
320      *
321      * @stable ICU 61
322      */
323     UNUM_SIGN_EXCEPT_ZERO,
324 
325     /**
326      * Use the locale-dependent accounting format on negative numbers, and show the plus sign on
327      * positive numbers. Do not show a sign on zero, numbers that round to zero, or NaN. For more
328      * information on the accounting format, see the ACCOUNTING sign display strategy.
329      *
330      * @stable ICU 61
331      */
332     UNUM_SIGN_ACCOUNTING_EXCEPT_ZERO,
333 
334     /**
335      * Same as AUTO, but do not show the sign on negative zero.
336      *
337      * @stable ICU 69
338      */
339     UNUM_SIGN_NEGATIVE,
340 
341     /**
342      * Same as ACCOUNTING, but do not show the sign on negative zero.
343      *
344      * @stable ICU 69
345      */
346     UNUM_SIGN_ACCOUNTING_NEGATIVE,
347 
348     // Do not conditionalize the following with #ifndef U_HIDE_INTERNAL_API,
349     // needed for unconditionalized struct MacroProps
350     /**
351      * One more than the highest UNumberSignDisplay value.
352      *
353      * @internal ICU 60: The numeric value may change over time; see ICU ticket #12420.
354      */
355     UNUM_SIGN_COUNT = 9,
356 } UNumberSignDisplay;
357 
358 /**
359  * An enum declaring how to render the decimal separator.
360  *
361  * <p>
362  * <ul>
363  * <li>UNUM_DECIMAL_SEPARATOR_AUTO: "1", "1.1"
364  * <li>UNUM_DECIMAL_SEPARATOR_ALWAYS: "1.", "1.1"
365  * </ul>
366  *
367  * @stable ICU 60
368  */
369 typedef enum UNumberDecimalSeparatorDisplay {
370     /**
371      * Show the decimal separator when there are one or more digits to display after the separator, and do not show
372      * it otherwise. This is the default behavior.
373      *
374      * @stable ICU 60
375      */
376             UNUM_DECIMAL_SEPARATOR_AUTO,
377 
378     /**
379      * Always show the decimal separator, even if there are no digits to display after the separator.
380      *
381      * @stable ICU 60
382      */
383             UNUM_DECIMAL_SEPARATOR_ALWAYS,
384 
385     // Do not conditionalize the following with #ifndef U_HIDE_INTERNAL_API,
386     // needed for unconditionalized struct MacroProps
387     /**
388      * One more than the highest UNumberDecimalSeparatorDisplay value.
389      *
390      * @internal ICU 60: The numeric value may change over time; see ICU ticket #12420.
391      */
392             UNUM_DECIMAL_SEPARATOR_COUNT
393 } UNumberDecimalSeparatorDisplay;
394 
395 /**
396  * An enum declaring how to render trailing zeros.
397  *
398  * - UNUM_TRAILING_ZERO_AUTO: 0.90, 1.00, 1.10
399  * - UNUM_TRAILING_ZERO_HIDE_IF_WHOLE: 0.90, 1, 1.10
400  *
401  * @stable ICU 69
402  */
403 typedef enum UNumberTrailingZeroDisplay {
404     /**
405      * Display trailing zeros according to the settings for minimum fraction and significant digits.
406      *
407      * @stable ICU 69
408      */
409     UNUM_TRAILING_ZERO_AUTO,
410 
411     /**
412      * Same as AUTO, but hide trailing zeros after the decimal separator if they are all zero.
413      *
414      * @stable ICU 69
415      */
416     UNUM_TRAILING_ZERO_HIDE_IF_WHOLE,
417 } UNumberTrailingZeroDisplay;
418 
419 struct UNumberFormatter;
420 /**
421  * C-compatible version of icu::number::LocalizedNumberFormatter.
422  *
423  * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
424  *
425  * @stable ICU 62
426  */
427 typedef struct UNumberFormatter UNumberFormatter;
428 
429 
430 /**
431  * Creates a new UNumberFormatter for the given skeleton string and locale. This is currently the only
432  * method for creating a new UNumberFormatter.
433  *
434  * Objects of type UNumberFormatter returned by this method are threadsafe.
435  *
436  * For more details on skeleton strings, see the documentation in numberformatter.h. For more details on
437  * the usage of this API, see the documentation at the top of unumberformatter.h.
438  *
439  * For more information on number skeleton strings, see:
440  * https://unicode-org.github.io/icu/userguide/format_parse/numbers/skeletons.html
441  *
442  * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
443  *
444  * @param skeleton The skeleton string, like u"percent precision-integer"
445  * @param skeletonLen The number of UChars in the skeleton string, or -1 if it is NUL-terminated.
446  * @param locale The NUL-terminated locale ID.
447  * @param ec Set if an error occurs.
448  * @stable ICU 62
449  */
450 U_CAPI UNumberFormatter* U_EXPORT2
451 unumf_openForSkeletonAndLocale(const UChar* skeleton, int32_t skeletonLen, const char* locale,
452                                UErrorCode* ec);
453 
454 
455 /**
456  * Like unumf_openForSkeletonAndLocale, but accepts a UParseError, which will be populated with the
457  * location of a skeleton syntax error if such a syntax error exists.
458  *
459  * For more information on number skeleton strings, see:
460  * https://unicode-org.github.io/icu/userguide/format_parse/numbers/skeletons.html
461  *
462  * @param skeleton The skeleton string, like u"percent precision-integer"
463  * @param skeletonLen The number of UChars in the skeleton string, or -1 if it is NUL-terminated.
464  * @param locale The NUL-terminated locale ID.
465  * @param perror A parse error struct populated if an error occurs when parsing. Can be NULL.
466  *               If no error occurs, perror->offset will be set to -1.
467  * @param ec Set if an error occurs.
468  * @stable ICU 64
469  */
470 U_CAPI UNumberFormatter* U_EXPORT2
471 unumf_openForSkeletonAndLocaleWithError(
472        const UChar* skeleton, int32_t skeletonLen, const char* locale, UParseError* perror, UErrorCode* ec);
473 
474 
475 
476 /**
477  * Uses a UNumberFormatter to format an integer to a UFormattedNumber. A string, field position, and other
478  * information can be retrieved from the UFormattedNumber.
479  *
480  * The UNumberFormatter can be shared between threads. Each thread should have its own local
481  * UFormattedNumber, however, for storing the result of the formatting operation.
482  *
483  * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
484  *
485  * @param uformatter A formatter object created by unumf_openForSkeletonAndLocale or similar.
486  * @param value The number to be formatted.
487  * @param uresult The object that will be mutated to store the result; see unumf_openResult.
488  * @param ec Set if an error occurs.
489  * @stable ICU 62
490  */
491 U_CAPI void U_EXPORT2
492 unumf_formatInt(const UNumberFormatter* uformatter, int64_t value, UFormattedNumber* uresult,
493                 UErrorCode* ec);
494 
495 
496 /**
497  * Uses a UNumberFormatter to format a double to a UFormattedNumber. A string, field position, and other
498  * information can be retrieved from the UFormattedNumber.
499  *
500  * The UNumberFormatter can be shared between threads. Each thread should have its own local
501  * UFormattedNumber, however, for storing the result of the formatting operation.
502  *
503  * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
504  *
505  * @param uformatter A formatter object created by unumf_openForSkeletonAndLocale or similar.
506  * @param value The number to be formatted.
507  * @param uresult The object that will be mutated to store the result; see unumf_openResult.
508  * @param ec Set if an error occurs.
509  * @stable ICU 62
510  */
511 U_CAPI void U_EXPORT2
512 unumf_formatDouble(const UNumberFormatter* uformatter, double value, UFormattedNumber* uresult,
513                    UErrorCode* ec);
514 
515 
516 /**
517  * Uses a UNumberFormatter to format a decimal number to a UFormattedNumber. A string, field position, and
518  * other information can be retrieved from the UFormattedNumber.
519  *
520  * The UNumberFormatter can be shared between threads. Each thread should have its own local
521  * UFormattedNumber, however, for storing the result of the formatting operation.
522  *
523  * The syntax of the unformatted number is a "numeric string" as defined in the Decimal Arithmetic
524  * Specification, available at http://speleotrove.com/decimal
525  *
526  * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
527  *
528  * @param uformatter A formatter object created by unumf_openForSkeletonAndLocale or similar.
529  * @param value The numeric string to be formatted.
530  * @param valueLen The length of the numeric string, or -1 if it is NUL-terminated.
531  * @param uresult The object that will be mutated to store the result; see unumf_openResult.
532  * @param ec Set if an error occurs.
533  * @stable ICU 62
534  */
535 U_CAPI void U_EXPORT2
536 unumf_formatDecimal(const UNumberFormatter* uformatter, const char* value, int32_t valueLen,
537                     UFormattedNumber* uresult, UErrorCode* ec);
538 
539 
540 
541 /**
542  * Releases the UNumberFormatter created by unumf_openForSkeletonAndLocale().
543  *
544  * @param uformatter An object created by unumf_openForSkeletonAndLocale().
545  * @stable ICU 62
546  */
547 U_CAPI void U_EXPORT2
548 unumf_close(UNumberFormatter* uformatter);
549 
550 
551 
552 #if U_SHOW_CPLUSPLUS_API
553 U_NAMESPACE_BEGIN
554 
555 /**
556  * \class LocalUNumberFormatterPointer
557  * "Smart pointer" class; closes a UNumberFormatter via unumf_close().
558  * For most methods see the LocalPointerBase base class.
559  *
560  * Usage:
561  * <pre>
562  * LocalUNumberFormatterPointer uformatter(unumf_openForSkeletonAndLocale(...));
563  * // no need to explicitly call unumf_close()
564  * </pre>
565  *
566  * @see LocalPointerBase
567  * @see LocalPointer
568  * @stable ICU 62
569  */
570 U_DEFINE_LOCAL_OPEN_POINTER(LocalUNumberFormatterPointer, UNumberFormatter, unumf_close);
571 
572 U_NAMESPACE_END
573 #endif // U_SHOW_CPLUSPLUS_API
574 
575 #endif /* #if !UCONFIG_NO_FORMATTING */
576 #endif //__UNUMBERFORMATTER_H__
577