Logo Search packages:      
Sourcecode: icu version File versions  Download package

SelectFormat Class Reference

#include <selfmt.h>

Inheritance diagram for SelectFormat:
Collaboration diagram for SelectFormat:

List of all members.

Public Member Functions

void applyPattern (const UnicodeString &pattern, UErrorCode &status)
virtual Formatclone (void) const
UnicodeStringformat (const Formattable &obj, UnicodeString &appendTo, UErrorCode &status) const
virtual UnicodeStringformat (const Formattable &obj, UnicodeString &appendTo, FieldPositionIterator *posIter, UErrorCode &status) const
UnicodeStringformat (const UnicodeString &keyword, UnicodeString &appendTo, FieldPosition &pos, UErrorCode &status) const
UnicodeStringformat (const Formattable &obj, UnicodeString &appendTo, FieldPosition &pos, UErrorCode &status) const
virtual UClassID getDynamicClassID () const
Locale getLocale (ULocDataLocaleType type, UErrorCode &status) const
const char * getLocaleID (ULocDataLocaleType type, UErrorCode &status) const
virtual UBool operator!= (const Format &other) const
SelectFormatoperator= (const SelectFormat &other)
virtual UBool operator== (const Format &other) const
void parseObject (const UnicodeString &source, Formattable &result, UErrorCode &status) const
virtual void parseObject (const UnicodeString &source, Formattable &result, ParsePosition &parse_pos) const
 SelectFormat (const UnicodeString &pattern, UErrorCode &status)
 SelectFormat (const SelectFormat &other)
UnicodeStringtoPattern (UnicodeString &appendTo)
virtual ~SelectFormat ()

Static Public Member Functions

static UClassID U_EXPORT2 getStaticClassID (void)
static void U_EXPORT2 operator delete (void *p) U_NO_THROW
static void U_EXPORT2 operator delete (void *, void *) U_NO_THROW
static void U_EXPORT2 operator delete[] (void *p) U_NO_THROW
static void *U_EXPORT2 operator new (size_t size) U_NO_THROW
static void *U_EXPORT2 operator new (size_t, void *ptr) U_NO_THROW
static void *U_EXPORT2 operator new[] (size_t size) U_NO_THROW

Protected Member Functions

void setLocaleIDs (const char *valid, const char *actual)

Static Protected Member Functions

static void syntaxError (const UnicodeString &pattern, int32_t pos, UParseError &parseError)

Static Private Member Functions

static int32_t findSubMessage (const MessagePattern &pattern, int32_t partIndex, const UnicodeString &keyword, UErrorCode &ec)

Private Attributes

MessagePattern msgPattern


class MessageFormat

Detailed Description

SelectFormat supports the creation of internationalized messages by selecting phrases based on keywords. The pattern specifies how to map keywords to phrases and provides a default phrase. The object provided to the format method is a string that's matched against the keywords. If there is a match, the corresponding phrase is selected; otherwise, the default phrase is used.

Using SelectFormat for Gender Agreement

Note: Typically, select formatting is done via MessageFormat with a select argument type, rather than using a stand-alone SelectFormat.

The main use case for the select format is gender based inflection. When names or nouns are inserted into sentences, their gender can affect pronouns, verb forms, articles, and adjectives. Special care needs to be taken for the case where the gender cannot be determined. The impact varies between languages:

  • English has three genders, and unknown gender is handled as a special case. Names use the gender of the named person (if known), nouns referring to people use natural gender, and inanimate objects are usually neutral. The gender only affects pronouns: "he", "she", "it", "they".
  • German differs from English in that the gender of nouns is rather arbitrary, even for nouns referring to people ("Mädchen", girl, is neutral). The gender affects pronouns ("er", "sie", "es"), articles ("der", "die", "das"), and adjective forms ("guter Mann", "gute Frau", "gutes Mädchen").
  • French has only two genders; as in German the gender of nouns is rather arbitrary - for sun and moon, the genders are the opposite of those in German. The gender affects pronouns ("il", "elle"), articles ("le", "la"), adjective forms ("bon", "bonne"), and sometimes verb forms ("allé", "allée").
  • Polish distinguishes five genders (or noun classes), human masculine, animate non-human masculine, inanimate masculine, feminine, and neuter.

Some other languages have noun classes that are not related to gender, but similar in grammatical use. Some African languages have around 20 noun classes.

Note:For the gender of a person in a given sentence, we usually need to distinguish only between female, male and other/unknown.

To enable localizers to create sentence patterns that take their language's gender dependencies into consideration, software has to provide information about the gender associated with a noun or name to MessageFormat. Two main cases can be distinguished:

  • For people, natural gender information should be maintained for each person. Keywords like "male", "female", "mixed" (for groups of people) and "unknown" could be used.

  • For nouns, grammatical gender information should be maintained for each noun and per language, e.g., in resource bundles. The keywords "masculine", "feminine", and "neuter" are commonly used, but some languages may require other keywords.

The resulting keyword is provided to MessageFormat as a parameter separate from the name or noun it's associated with. For example, to generate a message such as "Jean went to Paris", three separate arguments would be provided: The name of the person as argument 0, the gender of the person as argument 1, and the name of the city as argument 2. The sentence pattern for English, where the gender of the person has no impact on this simple sentence, would not refer to argument 1 at all:

{0} went to {2}.

Note: The entire sentence should be included (and partially repeated) inside each phrase. Otherwise translators would have to be trained on how to move bits of the sentence in and out of the select argument of a message. (The examples below do not follow this recommendation!)

The sentence pattern for French, where the gender of the person affects the form of the participle, uses a select format based on argument 1:

{0} est {1, select, female {allée} other {allé}} à {2}.

Patterns can be nested, so that it's possible to handle interactions of number and gender where necessary. For example, if the above sentence should allow for the names of several people to be inserted, the following sentence pattern can be used (with argument 0 the list of people's names, argument 1 the number of people, argument 2 their combined gender, and argument 3 the city name):

{0} {1, plural,
                 one {est {2, select, female {allée} other  {allé}}}
                 other {sont {2, select, female {allées} other {allés}}}
          }à {3}.

Patterns and Their Interpretation

The SelectFormat pattern string defines the phrase output for each user-defined keyword. The pattern is a sequence of (keyword, message) pairs. A keyword is a "pattern identifier": [^[[:Pattern_Syntax:][:Pattern_White_Space:]]]+

Each message is a MessageFormat pattern string enclosed in {curly braces}.

You always have to define a phrase for the default keyword other; this phrase is returned when the keyword provided to the format method matches no other keyword. If a pattern does not provide a phrase for other, the method it's provided to returns the error U_DEFAULT_KEYWORD_MISSING.
Pattern_White_Space between keywords and messages is ignored. Pattern_White_Space within a message is preserved and output.


 UErrorCode status = U_ZERO_ERROR;
 MessageFormat *msgFmt = new MessageFormat(UnicodeString("{0} est  {1, select, female {allée} other {allé}} à Paris."), Locale("fr"),  status);
 if (U_FAILURE(status)) {
 FieldPosition ignore(FieldPosition::DONT_CARE);
 UnicodeString result;

 char* str1= "Kirti,female";
 Formattable args1[] = {"Kirti","female"};
 msgFmt->format(args1, 2, result, ignore, status);
 cout << "Input is " << str1 << " and result is: " << result << endl;
 delete msgFmt;


Produces the output:
Kirti est allée à Paris.

ICU 4.4

Definition at line 183 of file selfmt.h.

The documentation for this class was generated from the following files:

Generated by  Doxygen 1.6.0   Back to index