For the random-by-key method of processing, a search argument that identifies the key of the record to be read is specified in factor 1 of the calculation specifications for the CHAIN operation. Figure 178 shows an example of an externally described DISK file being processed randomly by key. The specified record can be read from the file either during detail calculations or during total calculations.
The random-by-key method of processing is valid for a full procedural file designated as an input file or an update file.
For an externally described file, position 34 of the file description specification must contain a K, which indicates that the file is processed according to an access path that is built on keys.
The data description specifications (DDS) for the file specifies the field that contains the key value (the key field). Position 35 of the file description specification must be blank.
A program-described file must be designated as an indexed file (I in position 35), and position 34 of the file description specification must contain an A, D, G, P, T, or Z. The length of the key field is identified in positions 29-33 of the file description specification, and the starting location of the key field is specified on the KEYLOC keyword. Data description specifications must be used to create the access path for a program described input file (see Indexed File).