DB2 SQL Extension Toolkit - Release Notes

v1.3.1

April 20, 2020

• Fixed bug where the WORD_KEY and WORD_DIFF function was giving different outputs as compared to Netezza.
• Changes DDL for functions to NO EXTERNAL ACTION.

• Support for NVARCHAR/NCHAR functions added.
• README updated with the minimum CLPPLUS version.
• Fixed bug where either parameter of days_between is NULL.
• Fixed bug where renaming extracted directory leads to failure of registration.

• Migrated setup script from shell to Python for better user experience.
• Schema as an option is provided to register the UDXes.
• SSL connection using a DSN is now supported.
• Architecture check has been added which will notify the user if incorrect platform packages are used
• Added graceful error message for erroneous exit.
• Argument validation wherever applicable.

• Prefix DBTK_ has been removed from the functions' names.
• Max number of arguments accepted changed.
• Handled case where password contains special characters.
• Included README with the package.
• The project is also deleted now from the file-system after dropping all the UDXes while cleanup.

• For ARRAY_COMBINE function, more than one character delimiter does not work.
• ENCRYPT and COMPRESS results may vary as compared to PDA. If you encrypt a string using ENCRYPT function and use DECRYPT function on result of formerreturns original string. Similar for COMPRESS and DECOMPRESS functions. Eg. VALUES(DECRYPT(ENCRYPT('string')) = 'string'

• NVARCHAR is internally handled in VARCHAR in db2. Thus, registering the functions with the VARCHAR variant only, which will internally handle NVARCHAR as well.
• ARRAY objects are fit in VARCHAR column of the table. It is the user's responsibility to allocate appropriate memory in VARCHAR. Eg. VARCHAR(100) might give error while adding multiple BIGINT values. However, changing it to VARCHAR(10000) will solve the issue.
• Since BYTEINT is converted to SMALLINT in db2, ARRAY(1) gives an error for such a query. The user is advised to use ARRAY(2), as BYTEINT values are handled in SMALLINT.
• User needs to use functions with schema name under which they are registered. Schema will be the user name with which the package was installed.
• Output representation is different for DOUBLE/BOOLEAN in Db2 compared to PDA. So, you might observe output difference in case you are using DOUBLE/BOOLEAN data types with the functions.
• If your workload/query involves "NCHAR", "NVARCHAR" data then your system must be set with appropriate STRING_UNITS configurations.