Appendix A: Oracle Tuxedo Application Rehosting Workbench MBCS Support

This chapter contains the following topics:

•

Purpose

•	Procedures

•

Utility

Purpose

Oracle Tuxedo Application Rehosting Workbench (Tuxedo ART Workbench) supports Multiple-Byte Character Set (MBCS) for code conversion, which includes (but not limited to) Japanese, Chinese and Korean. Variable names with MBCS support for code conversion and MBCS support for data conversion are not supported.

Procedures

Converting the Code Containing the MBCS Characters

Following are the typical procedures to convert code which includes MBCS using Tuxedo ART Workbench:

1.	Fetch the source code assets (including COBOL, COPYBOOK, JCL, MAP) from mainframe using FTP tool with binary mode.

2.	Use code page conversion utility (e.g. ICU uconv) to convert the multiple-byte characters in EBCDIC encoding (e.g. IBM-1390) in the source assets to multiple-byte characters in encoding on open system (e.g.,. Shift-JIS).

3.	Use other tools provided by Tuxedo ART Workbench to convert the source code assets, e.g. COBOL Converter, JCL Converter.

Migrating the Data Containing the MBCS Characters

The typical procedures for migrating DB2 data (which includes MBCS characters) using Tuxedo ART Workbench are as follows:

1.	Set the following db-param.cfg parameters to generate the unload scripts which unloads the db2 data in “csv” format:

rdbms:jcl_unload_utility_name:dsnuproc

rdbms:jcl_unload_format_file:csv

2.	Invoke the Tuxedo ART Workbench on migration platform and upload the unload scripts into the source platform.

3.	Log in the source platform and unload DB2 data that contains MBCS characters.

4.	Run "tso ftp" command to connect the target platform.

5.	Set the transfer options using ftp commands:

locsite encoding=MBCS

locsite mbdataconn=(file_system_codepage, network_transfer_codep)

locsite mbsendeol=CRLF

file_system_codepage and network_transfer_codepagee are the corresponding Mainframe code pairs to your target MBCS. For example, for Simplified Chinese character set, choose IBM-5488 as file_system_codepage and IMB-1388 or UTF-8 as network_transfer_codepage.

6.	Run “ftp put”command to transfer the data file to target platform.

7.	Set NLS_LANG environment parameter and reload the data file without transcoding. For example, export NLS_LANG="SIMPLIFIED CHINESE_CHINA.XXXX"

For information about the typical procedure of converting the data file containing MBCS from DB2 on Mainframe, refer to the Tuxedo ART Workbench Reference Guide.

Utility

ICU uconv utility is recommended to convert the multiple-byte characters in EBCDIC encoding in the source assets to multiple-byte characters in encoding on open system.

The typical usage of uconv utility is as following:

uconv -f [source_encoding] -t [dest_encoding] -o [output_file_name] source_file_name

For example, the following command converts the input.bms file which is encoded in IBM-1390 to output.bms file which is encoded in Shift-JIS:

uconv -f ibm-1390 -t shift-jis -o output.bms input.bms