Setting a Multilingual Linguistic Sort

Use a multilingual linguistic sort if the database has data in:

• More than one language
• An Asian language

Multilingual linguistic sorts are useful for Asian languages because of their complex sorting rules.

If the data is in only one language, then use a monolingual linguistic sort to minimize memory use and to increase performance.

Multilingual linguistic sorts have three levels of precision:

• Primary level: A primary level sort distinguishes between base characters such as a and b.
• Secondary level: A secondary level sort distinguishes between base characters and also between the diacritical marks on a base character. For example, Ä and A are the same on the primary level because their base character is A, but they are different on the secondary level because of the diacritical mark on Ä.
• Tertiary level: A tertiary level sort distinguishes between base characters, between diacritical marks on a base character, and between the cases (upper and lower) of a character. For example, a and A are equal on the primary and secondary levels but different on the tertiary level because a is lowercase and A is uppercase.

Each locale must define the order of characters, that is, whether a is before b, b is before a, or they are equal.

You can choose a multilingual linguistic sort from a list of pre-defined sorts (called collations) or you can define a new sort. The pre-defined sorts are listed in the Existing Definitions dialog box that appears when you click Show Existing Definitions. Oracle Locale Builder allows you to view all the characteristics of any pre-defined multilingual linguistic sort.

You can define a new sort by modifying an existing sort or by entering all the data for a new sort. See Defining a New Multilingual Linguistic Sort.

Choosing the Collation Name from the Existing Definitions dialog box automatically completes the other field in the General multilingual sort definition dialog box:

• Collation ID: A numerical code that represents the chosen sort. It appears in parentheses after the Collation Name in the Existing Definitions dialog box

Choosing the collation name also determines whether one of the following Defined Collation Flags is checked:

• CANONICAL_EQUIVALENCE: Applies to Thai and Lao sorts
• REVERSE_SECONDARY: Applies to French sorts
• SWAP_WITH_NEXT: Determines whether canonical rules will be used

Defining a New Multilingual Sort

Each sort has several characteristics:

• Unicode collation sequence
• Non-spacing characters
• Punctuation characters
• Context-sensitive rules
• Expanding rules
• Base letters
• Special uppercase letters
• Special lowercase letters
• Special base letters

You can create a new multilingual linguistic sort by modifying an existing sort definition or by defining each sort characteristic. See Defining a New Multilingual Linguistic Sort.