The document topics covered by this document are as follows:
Adabas Native SQL V2.4.1 is compatible with Predict V4.5.x together with Predict V4.4, V4.3, V4.2 and V4.1.
Adabas Native SQL V2.4.1 is compatible with Adabas V7.1.x and above.
Note:
For information regarding Software AG product compatibility with IBM
platforms and any IBM requirements for Software AG products, please review the
Product Compatibility
for IBM Platforms web page.
For information on how long a product is supported by Software AG, access Software AG's Empower web site at https://empower.softwareag.com. In the left menu of this web page, expand Empower and log into Secured Services. Once you have logged in, you can expand Products in the left menu of the web page and select Product Roadmaps to access the Product Version Roadmaps application. This application allows you to request support information for specific products and releases.
Adabas V8.1 introduces new features and reduces some of the limits in previous versions. Adabas Native SQL V2.4.1 is adapted to support Adabas V8.1 with the following new features:
Native SQL will recognize the language specific generation code default in Predict "I8" and "R8" used to specify if descriptors will be generated in the Record buffer. (This is the same as it currently supports "I7" and "R7").
The new Large Object types of Adabas V8.1 will not be supported. However, the definition of Large object in Predict which is "LO" in the Field format definition will be recognized by Native SQL. In case of such field definition, Native SQL will issue a warning message and will skip this field.
The new ACBX calling mechanism will not be generated by Native SQL and there is a new error message in case Adabas buffers exceed the length of 32K.
The MU and PE occurrences for up to 65K will be supported, however Native SQL will keep the compatibility of previous versions:
The default of MU number of occurrences will be 191. This default will be used in case that no max occurrences is defined in Predict.
The default of PE number of occurrences will be 99. This default will be used in case that no max occurrences is defined in Predict.
The default of the counter field definition will be 2 binary bytes
field. However, there will be a new OPTIONS
parameter:
OPTIONS LONG-COUNTER
This will cause the generation of the counter field to be 4 binary bytes.
The 1-LAST definition for MU will generate 191 occurrences in the Record buffer. In case that the max occurrences number in Predict is larger than 191 occurrences, there will an error message and the record buffer generation will have only 191 occurrences.
The Format buffer and Search buffer contents will be generated with up to 5 digits for the occurrences number.
The internal field names of fields within the Value buffer in case of a specific occurrenc descriptor (which is part of a periodic group) will now have a suffix of up to 5 digits instead of two digits in previous versions.
The max size allowed for the generated Format buffer will be increased to 32K (instead of 10K). In Cobol a new redefined field will be generated on the Format Buffer Length field in the Adabas Control Block. In case of Format Buffer length greater that 9999 bytes, there will be an assignment of the real length to that field before the Adabas call.
Predict V4.5 is now recognized by Adabas native SQL including the new definitions for Adabas V8.1 features.
Native SQL is now supporting Unicode. The support is divided into two parts:
Alpha fields with Unicode format are defined in Predict as alphanumeric fields with SQL-CHAR-SET = "M". Native SQL checks every alphanumeric field for the SQL-CHAR-SET and in case of Unicode, it generates a "W" format in the Search and Format buffers. In addition the initial value generated in the Value buffer will be Low-values in Cobol and PL/1.
The CONNECT statement has now two new optional parameters to specify the encoding relevant to the session if different from the default set in the Adabas nucleus. The new parameters are:
[ACODE = value]
This value represents the key of supplied encoding for A fields during the user session.
[WCODE = value]
This value represents the key of supplied encoding for W fields during the user session.
In Cobol/V3 there is a compiler option "ARITH(EXTEND)" that
allows for numeric Unpacked and Packed fields to contain up to 31 digits.
Native SQL has a new OPTIONS parameter:
OPTIONS LARGE-NUMBERS
This parameter generates the Numeric and Packed fields in Cobol up to 31 digits. The user should be responsible for the appropriate compiler option in the Cobol compiler.
In case that the Predict file is defined under Adabas Vista
Configuration, the DBID and file number specified in the Native SQL
SYSFILE parameter do not match the physical Predict numbers that
are in the Predict Control record. With this version of Native SQL there is a
new OPTIONS parameter:
OPTIONS VISTA
This will cause Native SQL to ignore the unmatched numbers and take the numbers from the Control record to be used in the generation of the Global format id.
The RESPINT modules now display also the additions-2 field in case of an invalid Response code. The first two bytes as alphanumeric characters and the last two bytes as a binary number (the sub-code).
The following problems have been fixed in this new release:
P 286632 – Abend if the first field in the first SELECT is a PE field that is redefined.
P 274258– Wrong generation of a PE field in the first SELECT if followed by a field from another PE.
