FastSort Manual
Introduction to FastSort
FastSort Manual—429834-003
1-3
FastSort Components
Sequence options
Can use an alternate collating sequence for alphanumeric characters
Can remove records with duplicate key values
Input and output file options
Accepts up to 32 input Guardian files (except blocked tape files, SQL objects,
or processes)
Can merge up to 32 input streams from an application process
Supports application process input and output, with optional extended
addressing for sending and receiving records
Supports partitioned input and output files, including distributed Enscribe
databases (but not NonStop SQL/MP objects)
Produces relative, entry-sequenced, key-sequenced, or unstructured output
files
Supports both Format 1 and Format 2 files
Efficiency
Uses an optimized key-comparison algorithm and optimized procedures for
reading and building structured files, and double buffering for reading from
input files
Uses extended memory to minimize the number of merge passes
Automatically selects extended memory size
Uses record blocking to minimize the number of interprocess messages
required for transferring records to or from an application process
Supports parallel sorting to minimize sort time for files larger than 1 megabyte
Sorts files smaller than 100 kilobytes in memory to eliminate the need for
scratch file input and output
Provides record blocking and faster sorting for COBOL85, the File Utility
Program (FUP), and the CROSSREF program
Recognizes SORT and SUBSORT DEFINEs that allow you to control most
factors of a FastSort process from outside your sort program.
FastSort Components
FastSort has these major components:
Interactive FastSort
Programmatic FastSort