• Partitioned Data Sets (PDS) (Partitioned Data Set)
• Generation Data Group (GDG) (Generation Data Set Group),
• Sequential Data Sets (SEQ) (Sequential Data Set),
• Virtual Storage Access Method (VSAM) Data Sets(Virtual Sequential Access Method).The following sections discuss considerations you must take into account when configuring the Adapter for these data set types. This section also discusses the COBOL numeric data types the Adapter supports in COBOL Numeric Data Types.
• The names of the FilePublisher input, process, and output data sets must be a PDS name, for example TIBFA.TEST.IN.
• If more than one PDS file type is used in the configuration file, you should override the input, process, and output data sets within each file type with unique PDS names.
• Polling is only supported for PDS. FilePublisher uses the filePrefix parameter as a partial name, and checks the PDS for members that match the prefix. Any member name that matches the prefix is treated as an input file for publishing.
• File names that are passed in the trigger subject name are used as is. For example, for the most recent generation (0th generation) data file in a GDG, the file name can be specified as MYGDG(0). If no generation number is specified, the publisher reads the entire family of the GDG.
•
• The date and time are not appended to a final output file name by FileSubscriber (unlike Sequential file types).
• FileSubscriber does not use the filePrefix flag when writing VSAM files.
• FileSubscriber can only write data to an output VSAM file if the VSAM file has been allocated before subscribing.
• A VSAM file can only be published if the file name in the trigger message matches either the inputDataset or one of the vsamAltIndex filenames listed in the configuration file.The adapter supports the following COBOL numeric (non-text) data types. All types are supported by both publisher and subscriber.
• Zoned Decimal. In this type, the rightmost four bits of a byte are called the numeric bits (N) and normally consist of a code representing a decimal digit. The leftmost four bits of a byte are called the zone bits (Z), except for the rightmost byte of a decimal operand, where these bits may be treated either as a zone or as a sign (S).
• Binary (COMP and COMP-4). A binary type occupies 2, 4, or 8 bytes of storage and is handled for arithmetic purposes as a fixed-point number with the leftmost bit being the operational sign.
• Packed Decimal (COMP-3). In this type, each byte contains two decimal digits (D), except for the rightmost byte, which contains a sign to the right of a decimal digit.
• COMP-1. This is a 4-byte single-precision, floating point number. COMP-1 fields do not require a precision definition tag.
• COMP-2. This is an 8-byte double-precision, floating point number. COMP-2 fields do not require a precision definition tag.The precision of a field is defined as its length and number of decimals. Precision is specified for a field by using the “precision” tag. The format is precision="n,p" where n represents the number of digits for the number including decimals, and p represents the number of decimals.The length of a field in a record depends on two factors, the field’s numeric format and the precision specified.Binary Type Binary numbers occupy 2, 4, or 8 bytes of storage, depending on the precision specified.Packed Type Packed numbers occupy from 1 to 15 bytes of storage. The number of bytes is determined from the formula n/2 + 1. For example, a field with precision="7,2" would occupy 4 bytes of storage.When sending records from the Publisher with Packed Decimal fields (COMP-3), the message item description should set the type to either COMP-3 or PACKED. The packed data will be automatically converted to 8-byte floating point DOUBLE on the wire, unless the convertToString option is used. A convertToString="true" will cause the packed data to be converted into a DISPLAY format, sent as STRING on the wire. This avoids any loss of precision and rounding that can occur if FLOAT is used to represent decimal numbers. Sample definitions are shown below:When sending records with COMP-1 FLOAT, you may set the type field to COMP-1 or FLOAT. When sending COMP-2 DOUBLE, you may set the type field to COMP-2 or DOUBLE. When sending BINARY fields, you can either set the type field as SHORT (2 byte), INTEGER (4 byte) or LONGLONG (8 byte), or you may set the type field to BINARY or COMP, and explicitly set the length using the precision field as described above.When receiving records at the Subscriber, Packed Decimal field (COMP-3) message descriptions should match what was sent from the Publisher. If the Publisher sent the data on the wire as DOUBLE, the type field should be set to DOUBLE, and the convertTo option must be used, and set to COMP-3 or PACKED. If the Publisher sent the data on the wire as STRING, the type field should bet set to STRING, and the convertTo option must be used, and set to COMP-3 or PACKED. Sample definitions are shown below:
Copyright © TIBCO Software Inc. All Rights Reserved.