IBM DB2 Manual page 263

|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
to force a rollback operation and disables further operations on JDBC
connections that are in a unit of work during processing of JVM
shutdown hooks.
Any other value
The IBM Data Server Driver for JDBC and SQLJ takes no action with
respect to rollback processing during processing of JVM shutdown
hooks. This is the default.
db2.jcc.sendCharInputsUTF8
Specifies whether the IBM Data Server Driver for JDBC and SQLJ converts
character input data to the CCSID of the DB2 for z/OS database server, or
sends the data in UTF-8 encoding for conversion by the database server.
db2.jcc.sendCharInputsUTF8 applies to IBM Data Server Driver for JDBC and
SQLJ type 2 connectivity to DB2 for z/OS database servers only. If this
property is also set at the connection level, the connection-level setting
overrides this value.
Possible values are:
no, false, or 2
Specifies that the IBM Data Server Driver for JDBC and SQLJ converts
character input data to the target encoding before the data is sent to
the DB2 for z/OS database server. This is the default.
yes, true, or 1
Specifies that the IBM Data Server Driver for JDBC and SQLJ sends
character input data to the DB2 for z/OS database server in UTF-8
encoding. The data source converts the data from UTF-8 encoding to
the target CCSID.
Specify yes, true, or 1 only if conversion to the target CCSID by the
SDK for Java causes character conversion problems. The most common
problem occurs when you use IBM Data Server Driver for JDBC and
SQLJ type 2 connectivity to insert a Unicode line feed character
(U+000A) into a table column that has CCSID 37, and then retrieve that
data from a non-z/OS client. If the SDK for Java does the conversion
during insertion of the character into the column, the line feed
character is converted to the EBCDIC new line character X'15'.
However, during retrieval, some SDKs for Java on operating systems
other than z/OS convert the X'15' character to the Unicode next line
character (U+0085) instead of the line feed character (U+000A). The
next line character causes unexpected behavior for some XML parsers.
If you set db2.jcc.sendCharInputsUTF8 to yes, the DB2 for z/OS
database server converts the U+000A character to the EBCDIC line feed
character X'25' during insertion into the column, so the character is
always retrieved as a line feed character.
Conversion of data to the target CCSID on the data source might cause
the IBM Data Server Driver for JDBC and SQLJ to use more memory
than conversion by the driver. The driver allocates memory for
conversion of character data from the source encoding to the encoding
of the data that it sends to the data source. The amount of space that
the driver allocates for character data that is sent to a table column is
based on the maximum possible length of the data. UTF-8 data can
require up to three bytes for each character. Therefore, if the driver
sends UTF-8 data to the data source, the driver needs to allocate three
times the maximum number of characters in the input data. If the
Chapter 7. JDBC and SQLJ reference information
247