Resolving Data Retrieval Problems

This section of the Readme contains known data retrieval (Universal Data Access) issues that apply to several components of the solution.

Large Memory Demands When Returning Large Results Sets Using Certain Versions of Oracle 9iR2 Client on UNIX

When running queries that return large result sets, unexpectedly large amounts of memory are consumed on UNIX systems when using certain versions of the Oracle 9iR2 client software. For example, when returning 5,000 or more rows using Oracle 9iR2 client software that is earlier than version 9.2.0.5, some reports may take 150% to 500% of the memory typically used by later versions of this client or the Oracle 8i software.

To resolve this problem, we recommend that you use version 9.2.0.5 or higher versions of the Oracle UNIX database client software.

439963, 443483

Support for Japanese Vendor-Defined Characters in IBM Cognos Series 7

IBM Cognos Series 7 includes several mechanisms that manage the Japanese vendor-defined characters (VDC) that were assigned with multiple code points in the Shift-JIS encoding standard. These characters are allocated in the following code ranges:

• NEC Selected Special Symbols: 0x8740 - 0x879C
• NEC Selected IBM Extensions: 0xED40 - 0xEEFC
• IBM Extensions: 0xFA40 - 0xFC4B

Characters and symbols that are assigned with multiple code points are treated as restricted characters. A support mechanism exists for the following restricted characters.

Kanji Restricted Characters 

IBM - 0xFA5C – 0xFC4B are the supported code points for these multiple code points:

• IBM - 0xFA5C - 0xFC4B

• NEC - 0xED40 - 0xEEEC

Small Roman Numeral Restricted Characters

For the following multiple code points, IBM - 0xFA40 - 0xFA49 are the supported code points:

• IBM - 0xFA40 - 0xFA49
• NEC - 0xEEEF - 0xEEF8

Roman Numeral Restricted Characters

For the following multiple code points, NEC - 0x8754 - 0x875D are the supported code points:

• IBM - 0xFA4A - 0xFA53
• NEC - 0x8754 - 0x875D

Parenthesized Ideograph Stock (Kabu ) Restricted Characters

For any multiple code point combination, 0x878A is the supported code point. 

Numero Sign Restricted Character

For any multiple code point combination, 0x8782 is the supported code point.

Telephone Sign Restricted Character

For any multiple code point combination, 0x8784 is the supported code point.

Fullwidth Not Sign Restricted Character

For any multiple code point combination, 0x81CA is the supported code point.

Fullwidth Broken Bar Restricted Character

For any multiple code point combination, 0xFA55 is the supported code point.

Fullwidth Apostrophe Restricted Character

For any multiple code point combination, 0xFA56 is the supported code point.

Fullwidth Quotation Mark Restricted Character

For any multiple code point combination, 0xFA57 is the supported code point.

Because Restricted Character

For any multiple code point combination, 0x81E6 is the supported code point.

Other Restricted Characters

For the following multiple code points, regular symbol - 0x8140 - 0x81FC is the supported code point:

• regular symbol - 0x8140 - 0x81FC
• NEC - 0x8790 - 0x879C

Non-Restricted Characters

All VDC characters and symbols that are assigned with a single code point are treated as non-restricted characters and they are supported as other Japanese characters.

Unsupported Characters

IBM Cognos Series 7 doesn’t supply a mechanism to detect or warn users when non-supported code points are entered.

Additional Japanese Encoding Support

In addition to the Japanese Shift-JIS encoding, IBM Cognos NoticeCast supports EUC-JP and ISO-2022-JP encoding for outgoing email. This feature will allow the customer to read alerts and notifications sent by email systems that require these encodings. However, Japanese VDC code ranges, especially restricted characters, are not fully supported by JAVA runtime. When IBM Cognos NoticeCast is configured to use EUC-JP or ISO-2022-JP as the encoding for outgoing email, using the VDC characters in email body and subject are not supported.

nbna