Storing Foreign Characters in oracle 9i Lite Client Database.

Question

Hi All,
My Database configuration Is
RDBMS VERSION 9.0.1.1.1
NLS_CHARACTERSET AL32UTF8
NLS_NCHAR_CHARACTERSET AL16UTF16
NLS_LANGUAGE AMERICAN
NLS_TERRITORY AMERICA
Operating System Windows 2000
My Oracle Lite configuration Is
Oracle9i Lite Release 5.0.2.0.8
Client Locations:
Africa,Europe,Asia,America
Currently all clients store data only in English in the client odb either through java and oracle forms applications.
We want to now store data for some specific tables, in the clients own language.Also we should be able to retrieve that data and see that data as it was entered.
For example :
The client in china after downloading the odb should be able to enter the data for that specific tables in the Chinese language.(They use windows 2000/XP English version).
1. What are the settings i need to do in the client environment?
like installing Chinese font or setting NLS_LANG parameter etc...
2. What are the settings i need to do in the server environment?
Because there should not be any loss of any data or junk values getting stored instead of actual values entered by clients.
Also when i query that specific tables in the oracle database after synchronization i should be able to view that data entered by the clients in their own language.
Eg:data entered by china clients should be visible in chinese fonts,
data entered by brazil clients should be visible in brazilian fonts,
data entered by vietnam clients should be visible in vietnamese fonts.
3. will there be any synchronization issues when clients try to sync after data entry with
their own languages?
Please do advice on these issue...........

Answer 1

I have given you the info of how to set up Oracle Lite for UTF8. I can only point you in the direction for globalization on the enterprise server. Also note that Oracle Lite only support NLS_LENGHT_SEMANTICS=BYTE, your EE database will be CHAR, so you have to alter the session paramter before you create the repository and/or publish your application.
Here is a Oracle Document on Globalization:
http://www.oracle.com/technology/tech/globalization/pdf/TWP_AppDev_Unicode_10gR2.pdf
Here is what Oracle Lite's Developers Guide has on development of linguistic sorts:
2.11 Support for Linguistic Sort
Linguistic sort is a feature for the ASCII version of Oracle Database Lite. It produces culturally acceptable order of strings for a specified language or collation sequence. The ASCII version supports several code pages defined by single-byte 8-bit encoding schemes. Each of these code pages is a super set of 7-bit ASCII, and the additional accented characters necessary to support a group of European languages are included in the upper 128 bytes. A new string comparison mechanism is provided that produces strings in a linguistically correct order by mapping each collation element of a string to the corresponding 8-bit value of the supported code page.
2.11.1 Creating Linguistic Sort Enabled Databases
The linguistic sort capability must be enabled when the database is created using the CREATEDB command line utility with the <collation_sequence> enabled.
Note:
For more information on the CREATEDB utility, see Section A.2, "CREATEDB".
The behavior of the ORDER_BY clause and the WHERE condition are determined by how the NLS_SORT parameter is implemented. Binary sorting is the default setting, and is used unless the <collation_sequence> parameter is set to use the linguistic sort ordering rules.
NLSRT is not supported in the current version of Oracle Database Lite. Therefore, NCHAR data type is not yet available.
2.11.2 How Collation Works
Collation refers to ordering of strings into a culturally acceptable sequence. A collation sequence is a sequence of all collation elements from an alphabet from smallest collation order to the largest. Once a collation sequence is given, orders of all strings from the same alphabet are fixed. As such, the collation sequence encodes the linguistic requirements on collation. A collation element is the smallest sub-string that can be used by the comparison function to determine the order of two strings.
2.11.3 Collation Element Examples
Normally, a collation element is just one character. In binary sorting, only one property, the code value that represents a character, is used. But in linguistic sorting, usually three properties. The primary level of difference is the base character. The secondary level of difference is for diacritical marks on a given base character. The tertiary level of difference is for the case of a given character. Punctuation can function as a fourth level of difference, but comparisons for punctuation occur last and are made at the binary rather than the linguistic level. These are used for each collation element. The following sections contain examples that demonstrate sorting priorities.
2.11.3.1 Sorting Normal Characters
This section lists a set of examples that describe how to sort normal characters.
Example 1
'a' < 'b'. There is a primary difference between them on the character level.
Example 2
'À' > 'a'. This difference occurs on the secondary level. Note that 'À'and 'a' are considered "equal" on the primary level.
Example 3
'À' < 'à' in FRENCH but 'À' > 'à' in GERMAN. This difference on the tertiary level. Note that 'À' and 'à' are considered being "equal" on the primary and secondary level. Also note that the case convention may be different for different language.
Example 4
'às' < 'at'. This is a difference on the primary level. This example shows the role of difference levels: the lower level differences are ignored if there is a primary level difference anywhere in the strings.
Example 5
'+data' < '-data' <'data' <'data-'. If strings are compared and present no difference on the primary, secondary, or tertiary levels, they are compared for punctuation.
2.11.3.2 Reverse Sorting of French Accents
Some languages, particularly French, require words to be ordered on the secondary level according to the last accent difference. This behavior is known as French secondary sorting or French accent ordering.
Example
'côte' < 'coté' in FRENCH but 'coté' < 'côte' in GERMAN. Note that the secondary difference of 'e' and 'é' occurred later than those of 'ô' and 'o'.
2.11.3.3 Sorting Contracting Characters
There are some special cases where two or more characters in a group can function as a single collation element. These types of collation elements are called 'contracting characters' or 'group characters'. In these cases each of these characters properties are assigned appropriate values.
Example
'h' < 'ch' < 'i' in XCZECH. Here 'ch' is assigned a primary property value which differentiates it from 'h' and 'i', such that 'h' < 'ch' < 'i'. Note that 'ch' is treated as a single character.
2.11.3.4 Sorting Expanding Characters
If a letter sorts as if it were a sequence of more than one letter, it is called an 'expanding character'. For example, in German the sharp s (ß) is treated as if it were a string of two characters 'ss' when comparing with other letters.
2.11.3.5 Sorting Numeric Characters
Only sorting of single digit characters from '0' to '9' is currently supported. For the supported European languages a digit character is always sorted as greater than any alphabetic character. For other languages this may be not the same. Other numeric characters such as Roman numeric characters and counting sequences, such as "one", "two", "three", are not supported at this time.
Example
'1' > 'z' in any European language, '1' < 'a' in LATVIAN. Note that this difference occurs on the primary level.