Neoview Character Sets Administrator's Guide (R2.3)

Table 3-1 Summary of SQL Language Rules by Neoview Character Set Configuration (continued)
Unicode ConfigurationSJIS ConfigurationISO88591 ConfigurationSQL Language Rule
Use these prefixes:
• In UCS2 string literals, you
can but are not required to
specify:
— _UCS2
for UCS2 characters or
— N for NCHAR, which
maps by default to
UCS2 characters
• In ISO8859-1 string literals,
you must specify:
— _ISO88591, which
maps to UTF8
characters
Use these prefixes:
• In ISO8859-1 string
literals, you can but are
not required to specify:
— _ISO88591 for SJIS
characters
• In UCS2 string literals,
you must specify:
— _UCS2
for UCS2 characters or
— N for NCHAR, which
maps by default to
UCS2 characters
Use these prefixes:
• In ISO8859-1 string
literals, you can but are
not required to specify:
— _ISO88591 for
ISO8859-1 characters
• In UCS2 string literals,
you must specify:
— _UCS2
for the UCS2 character
set or
— N for NCHAR, which
maps by default to
UCS2 characters
Explicitly specify a
valid character set
prefix value
(_character-set)
for every string literal
in a column that is not
in the default column
for the configuration.
SQL Functions and SQL String Functions
In the Unicode configuration:
• For UCS2 data, string
functions work on two-byte
character boundaries. For
ISO8859-1 data, string
functions work on
single-byte character
boundaries.
• Surrogate pairs are treated
as 2-byte characters.
In the SJIS configuration:
• For ISO8859-1 data, the
SQL string functions work
on bytes, not on SJIS
characters. As a result,
characters might be
truncated or might not be
found. For information
about troubleshooting
truncated characters in a
character string literal, see
Table 5-1 (page 35).
• For ISO8859-1 data,
substring functions return
single-byte characters. For
UCS2 data, substring
functions return two-byte
characters.
In the ISO88591
configuration:
• For ISO8859-1 character
data, string functions
work on single-byte
character boundaries. For
UCS2 data, string
functions work on
two-byte character
boundaries.
• For ISO8859-1 character
data, substring functions
return single-byte
characters. For UCS2 data,
substring functions return
two-byte characters.
The behavior of SQL
functions and SQL
string functions is
determined by the
selected Neoview
character set
configuration. For
information about SQL
functions, see Table 3-2
(page 26). For more
information about SQL
string functions, see
Table 3-3 (page 26).
Binary Collations
For UCS2 columns, binary
collations are based on UCS2
encoding. For ISO88591
columns. binary collations are
based on UTF8 encoding.
For ISO88591 columns,
binary collations are based
on SJIS encoding. For UCS2
columns, binary collations
are based on UCS2 encoding.
For ISO88591 columns,
binary collations are based
on ISO8859-1 encoding. For
UCS2 column, binary
collations are based on UCS2
encoding.
The type of encoding
used in binary
collations is
determined by the
selected configuration.
User-Defined Characters
User-defined GB2312 and
GBK characters can be stored
either in ISO88591 columns
(encoded in UTF8) or in UCS2
columns (encoded in UCS2).
User-defined SJIS characters
can be stored in ISO88591
columns.
Binary data can be stored in
ISO88591 columns. Such data
can include user-defined
characters.
Provide user-defined
characters only as
permitted by the
selected configuration.
SQL Identifiers
24 Using SQL Language Elements to Define and Manage Database Encoding