1. Oracle Text Reference
  2. Oracle Text Multilingual Features

D Oracle Text Multilingual Features

This Appendix describes the multilingual features of Oracle Text. The following topics are discussed:

D.1 Introduction

This appendix summarizes the main multilingual features for Oracle Text.

For a complete list of Oracle Globalization Support languages and character set support, refer to the Oracle Database Globalization Support Guide.

Note:

Oracle Text does not support the NLS_COMP and NLS_SORT parameters. Search results generated from Oracle Text are independent from values of those parameters.

In Oracle Database 12c Release 2 (12.2), an Oracle Text index cannot be created on a column with a declared collation other than BINARY, USING_NLS_COMP, USING_NLS_SORT or USING_NLS_SORT_CS. For all the supported collations, the Oracle Text behavior is the same.

D.2 Indexing

The following sections describe the multilingual indexing features:

D.2.1 Multilingual Features for Text Index Types

The following sections describes the supported multilingual features for the Oracle Text index types.

See Also:

"Lexer Types" for a description of available lexers

D.2.1.1 CONTEXT Index Type

The CONTEXT index type fully supports multilingual features, including use of the language and character set columns.

The following lexers are supported:

  • AUTO_LEXER

  • BASIC_LEXER

  • MULTI_LEXER

  • USER_LEXER

  • WORLD_LEXER

  • CHINESE_LEXER

  • CHINESE_VGRAM_LEXER

  • JAPANESE_LEXER

  • JAPANESE_VGRAM_LEXER

  • KOREAN_MORPH_LEXER

D.2.1.2 CTXCAT Index Type

CTXCAT supports the multilingual features of the BASIC_LEXER with the exception of indexing themes, and supports the following additional lexers:

  • USER_LEXER

  • WORLD_LEXER

CTXCAT also supports the following lexers:

  • CHINESE_LEXER

  • CHINESE_VGRAM_LEXER

  • JAPANESE_LEXER

  • JAPANESE_VGRAM_LEXER

  • KOREAN_MORPH_LEXER

Note:

The Oracle Text indextype CTXCAT is deprecated with Oracle Database 23ai. The indextype itself, and it's operator CTXCAT, can be removed in a future release.

Both CTXCAT and the use of CTXCAT grammar as an alternative grammar for CONTEXT queries is deprecated. Instead, Oracle recommends that you use the CONTEXT indextype, which can provide all the same functionality, except that it is not transactional. Near-transactional behavior in CONTEXT can be achieved by using SYNC(ON COMMIT) or, preferably, SYNC(EVERY [time-period]) with a short time period.

CTXCAT was introduced when indexes were typically a few megabytes in size. Modern, large indexes, can be difficult to manage with CTXCAT. The addition of index sets to CTXCAT can be achieved more effectively by the use of FILTER BY and ORDER BY columns, or SDATA, or both, in the CONTEXT indextype. CTXCAT is therefore rarely an appropriate choice. Oracle recommends that you choose the more efficient CONTEXT indextype.

D.2.1.3 CTXRULE Index Type

The CTXRULE index type supports the multilingual features of the BASIC_LEXER including ABOUT and STEM operators. It also supports Japanese, Chinese, and Korean (when used with the SVM_CLASSIFIER).

D.2.2 Lexer Types

Oracle Text supports the indexing of different languages by enabling you to choose a lexer in the indexing process. The specified lexer determines the languages you can index.

Table D-1 Oracle Text Lexer Types

Lexer Supported Languages

AUTO_LEXER

Automatically identifies the language being indexed by examining the content, and applies suitable options (including stemming) for that language. Works best where each document contains a single-language, and has at least a couple of paragraphs of text to aid identification.

BASIC_LEXER

Extracts tokens from text in languages, such as English and most of the western European languages that use space-delimited words.

MULTI_LEXER

Indexes tables containing documents of different languages such as English, German, and Japanese.

CHINESE_VGRAM

Extracts tokens from Chinese text.

CHINESE_LEXER

Extracts tokens from Chinese text. This lexer offers the following benefits over the CHINESE_VGRAM lexer:

  • Generates a smaller index

  • Better query response time

  • Generates real world tokens resulting in better query precision

  • Supports stop words

JAPANESE_VGRAM

Extracts tokens from Japanese text.

JAPANESE_LEXER

Extracts tokens from Japanese text. This lexer offers the following advantages over the JAPANESE_VGRAM lexer:

  • Generates smaller index

  • Better query response time

  • Generates real world tokens resulting in better precision

KOREAN_MORPH_LEXER

Extracts tokens from Korean text.

USER_LEXER

Indexes a particular language.

WORLD_LEXER

Indexes tables containing documents of different languages; autodetects languages in a document

D.2.3 Basic Lexer Features

The following features are supported with the BASIC_LEXER preference. Enable these features with attributes of the BASIC_LEXER. Features such as alternate spelling, composite, and base letter can be enabled together for better search results.

D.2.3.1 Theme Indexing

Enables the indexing and subsequent querying of document concepts with the ABOUT operator with CONTEXT index types. These concepts are derived from the Oracle Text knowledge base. This feature is supported for English and French.

This feature is not supported with CTXCAT index types.

D.2.3.2 Alternate Spelling

This feature enables you to search on alternate spellings of words. For example, with alternate spelling enabled in German, a query on gross returns documents that contain groß and gross.

This feature is supported in German, Danish, and Swedish.

Additionally, German can be indexed according to both traditional and reformed spelling conventions.

See Also:

"Alternate Spelling" and "New German Spelling".

D.2.3.3 Base Letter Conversion

This feature enables you to query words with or without diacritical marks such as tildes, accents, and umlauts. For example, with a Spanish base-letter index, a query of energia matches documents containing both energía and energia.

This feature is supported for English and all other supported whitespace delimited languages. In English and French, you can use the basic lexer to enable theme indexing.

See Also:

"Base-Letter Conversion"

D.2.3.4 Composite

This feature enables you to search for words that contain the specified term as a sub-composite. You must use the stem ($) operator.

For example, in German, a query of $register finds documents that contain Bruttoregistertonne and Registertonne.

You can use this feature for all languages that are supported for the INDEX_STEMS attribute of BASIC_LEXER.

D.2.3.5 Index Stems

This feature enables you to specify a stemmer for stem indexing.

Tokens are stemmed to a single base form at index time in addition to the normal forms. Specifying index stems enables better query performance for stem queries, for example $computed.

You can use this feature for all languages that are supported for the INDEX_STEMS attribute of BASIC_LEXER.

D.2.4 Multi Lexer Features

The MULTI_LEXER lexer enables you to index a column that contains documents of different languages. During indexing Oracle Text examines the language column and switches in the language-specific lexer to process the document. Define the lexer preferences for each language before indexing.

The multi lexer enables you to set different preferences for languages. For example, you can have composite set to TRUE for German documents and composite set to FALSE for Dutch documents.

D.2.5 World Lexer Features

Like MULTI_LEXER, the WORLD_LEXER lexer enables you to index documents that contain different languages. It automatically detects the languages of a document and, therefore, does not require you to create a language column in the base table.

WORLD_LEXER processes all database character sets and supports the Unicode 5.0 standard. For WORLD_LEXER to be effective with documents that use multiple languages, AL32UTF-8 or UTF8 Oracle character set encoding must be specified. This includes supplementary, or "surrogate-pair," characters.

Table D-2 and Table D-3 show the languages supported by WORLD_LEXER. This list may change as the Unicode standard changes, and in any case should not be considered exhaustive. (Languages are grouped by Unicode writing system, not by natural language groupings.)

Table D-2 Languages Supported by the World Lexer (Space-separated)

Language Group Languages Include

Arabic

Arabic, Farsi, Kurdish, Pashto, Sindhi, Urdu

Armenian

Armenian

Bengali

Assamese, Bengali

Bopomofo

Hakka Chinese, Minnan Chinese

Cyrillic

Over 50 languages, including Belorussian, Bulgarian, Macedonian, Moldavian, Russian, Serbian, Serbo-Croatian, Ukrainian

Devenagari

Bhojpuri, Bihari, Hindi, Kashmiri, Marathi, Nepali, Pali, Sanskrit

Ethiopic

Amharic, Ge'ez, Tigrinya, Tigre

Georgian

Georgian

Greek

Greek

Gujarati

Gujarati, Kacchi

Gurmukhi

Punjabi

Hebrew

Hebrew, Ladino, Yiddish

Kaganga

Redjang

Kannada

Kanarese, Kannada

Korean

Korean, Hanja Hangul

Latin

Afrikaans, Albanian, Basque, Breton, Catalan, Croatian, Czech, Danish, Dutch, English, Esperanto, Estonian, Faeroese, Fijian, Finnish, Flemish, French, Frisian, German, Hawaiian, Hungarian, Icelandic, Indonesian, Irish, Italian, Lappish, Classic Latin, Latvian, Lithuanian, Malay, Maltese, Pinyin Mandarin, Maori, Norwegian, Polish, Portuguese, Provencal, Romanian, Rumanian, Samoan, Scottish Gaelic, Slovak, Slovene, Slovenian, Sorbian, Spanish, Swahili, Swedish, Tagalog, Turkish, Vietnamese, Welsh

Malayalam

Malayalam

Mongolian

Mongolian

Oriya

Oriya

Sinhalese, Sinhala

Pali, Sinhalese

Syriac

Aramaic, Syriac

Tamil

Tamil

Telugu

Telugu

Thaana

Dhiveli, Divehi, Maldivian

Table D-3 Languages Supported by the World Lexer (Non-space-separated)

Language Group Languages Include

Chinese

Cantonese, Mandarin, Pinyin phonograms

Japanese

Japanese (Hiragana, Kanji, Katakana)

Khmer

Cambodian, Khmer

Lao

Lao

Myanmar

Burmese

Thai

Thai

Tibetan

Dzongkha, Tibetan

Table D-4 shows languages not supported by the World Lexer.

Table D-4 Languages Not Supported by the World Lexer

Language Group Languages Include

Buhid

Buhid

Canadian Syllabics

Blackfoot, Carrier, Cree, Dakhelh, Inuit, Inuktitut, Naskapi, Nunavik, Nunavut, Ojibwe, Sayisi, Slavey

Cherokee

Cherokee

Cypriot

Cypriot

Limbu

Limbu

Ogham

Ogham

Runic

Runic

Tai Le (Tai Lu, Lue, Dai Le)

Tai Le

Ugaritic

Ugaritic

Yi

Yi

Yi Jang Hexagram

Yi Jang

D.3 Querying

Oracle Text supports the use of different query operators. Some operators can be set to behave in accordance with your language. This section summarizes the multilingual query features for these operators.

  • Use the ABOUT operator to query on concepts. The system looks up concept information in the theme component of the index. This feature is supported for English and French with CONTEXT indexes only.

  • The fuzzy operator enables you to search for words that have similar spelling to specified word. Oracle Text supports fuzzy for English, French, German, Italian, Dutch, Spanish, Portuguese, Japanese, Optical Character recognition (OCR), and automatic language detection.

  • The stem operator enables you to search for words that have the same root as the specified term. For example, a stem of $sing expands into a query on the words sang, sung, sing. The Oracle Text stemmer supports the following languages: English, French, Spanish, Italian, German, Japanese and Dutch.

D.4 Supplied Stoplists

By default, the system indexes text using the Oracle Text supplied stoplists that correspond to your database language.

A stoplist is a list of stopwords that do not get indexed. These are usually common words in a language, such as this, that, and can in English. By default, all such words are defined in the Oracle Text supplied stoplists. You can customize these stoplists or update the stopwords based on your requirements.

Supported Languages and Stoplists Location

The Oracle Text supplied stoplists contain a list of stopwords, which are provided as defaults for all BASIC_LEXER and AUTO_LEXER supported languages. These stopwords are automatically loaded during installation or upgrade for the chosen database language.

The default stoplists (along with other default preferences) are defined in the administration (SQL) files, which are located in the $ORACLE_HOME/ctx/admin directory. These SQL files are named drdefLANG.sql, where LANG specifies the language code. For example, the default stoplist for French (language code: f) is defined in the $ORACLE_HOME/ctx/admin/drdeff.sql file.

The source files for these default stoplists contain a list of stopwords, and are located in the $ORACLE_HOME/ctx/data/stoplist directory. These source files are named drstopLANG.txt, where LANG specifies the language code. The contents of the source files are the extracted terms from the drdefLANG.sql files.

For a list of all languages (and their language codes) in which default stoplists are supplied, see Multilingual Features Matrix.

How to Load Your Own Stoplists

By default, only one drdefLANG.sql file is loaded during installation or upgrade based on the database language that you choose. You can call the CTX_DDL.LOAD_STOPLIST procedure to customize your stoplist or modify the default list of stopwords.

Unlike CTX_DDL.ADD_STOPWORD (which adds a single stopword per call), CTX_DDL.LOAD_STOPLIST takes a source file of stopwords for your specified language (from $ORACLE_HOME/ctx/data/stoplist/drstopLANG.txt) and loads to your stoplist.

Related Topics

D.5 Knowledge Base

An Oracle Text knowledge base is a hierarchical tree of concepts used for theme indexing, ABOUT queries, and deriving themes for document services.

Oracle Text supplies knowledge bases in English and French only. These knowledge bases are installed by default.

You can extend theme functionality to languages other than English or French by loading your own knowledge base for any single byte white space delimited language, including Spanish.

D.6 Multilingual Features Matrix

These are the multilingual features for all supported languages.

Table D-5 Multilingual Features for Supported Languages

Language Name Language Code Alternate Spelling Fuzzy Matching Language-Specific Lexer Default Stoplist Stemming

Afrikaans

af

N/A

No

Yes

Yes

Yes

Arabic

ar

N/A

No

Yes

Yes

Yes

Basque

eu

N/A

No

Yes

Yes

Yes

Belarusian

be

N/A

No

Yes

Yes

Yes

Bokmal (Norwegian)

n

N/A

No

Yes

Yes

Yes

Bulgarian

bg

N/A

No

Yes

Yes

Yes

Catalan

ca

N/A

No

Yes

Yes

Yes

Simplified Chinese

zh-cn

N/A

No

Yes

Yes

Yes

Croatian

hr

N/A

No

Yes

Yes

Yes

Czech

cs

N/A

No

Yes

Yes

Yes

Danish

dk

Yes

No

Yes

Yes

Yes

Dutch

nl

N/A

Yes

Yes

Yes

Yes

English

us

N/A

Yes

Yes

Yes

Yes

Estonian

et

N/A

No

Yes

Yes

Yes

Finnish

sf

N/A

No

Yes

Yes

Yes

French

f

N/A

Yes

Yes

Yes

Yes

Galician

gl

N/A

No

Yes

Yes

Yes

German

d

Yes

Yes

Yes

Yes

Yes

Greek

el

N/A

No

Yes

Yes

Yes

Hebrew

iw

N/A

No

Yes

Yes

Yes

Hindi

hi

N/A

No

Yes

Yes

Yes

Hungarian

hu

N/A

No

Yes

Yes

Yes

Icelandic

is

N/A

No

Yes

Yes

Yes

Indonesian

in

N/A

No

Yes

Yes

Yes

Italian

i

N/A

Yes

Yes

Yes

Yes

Japanese

ja

N/A

Yes

Yes

Yes

Yes

Korean

ko

N/A

No

Yes

Yes

Yes

Latvian

lv

N/A

No

Yes

Yes

Yes

Lithuanian

lt

N/A

No

Yes

Yes

Yes

Macedonian

mk

N/A

No

Yes

Yes

Yes

Malay

ms

N/A

No

Yes

Yes

Yes

Nynorsk (Norwegian)

nn

N/A

No

Yes

Yes

Yes

Persian (Farsi)

fa

N/A

No

Yes

Yes

Yes

Polish

pl

N/A

No

Yes

Yes

Yes

Portuguese

pt

N/A

Yes

Yes

Yes

Yes

Romanian

ro

N/A

No

Yes

Yes

Yes

Russian

ru

N/A

No

Yes

Yes

Yes

Slovak

sk

N/A

No

Yes

Yes

Yes

Slovenian

sl

N/A

No

Yes

Yes

Yes

Serbian

sr

N/A

No

Yes

Yes

Yes

Spanish

es

N/A

Yes

Yes

Yes

Yes

Swedish

sv

Yes

No

Yes

Yes

Yes

Thai

th

N/A

No

Yes

Yes

Yes

Traditional Chinese

zh-tw

N/A

No

Yes

Yes

Yes

Turkish

tr

N/A

No

Yes

Yes

Yes

Ukrainian

uk

N/A

No

Yes

Yes

Yes

Urdu

ur

N/A

No

Yes

Yes

Yes