Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Record-Type Change Method by Binary Relations

IP.com Disclosure Number: IPCOM000101246D
Original Publication Date: 1990-Jul-01
Included in the Prior Art Database: 2005-Mar-16
Document File: 3 page(s) / 91K

Publishing Venue

IBM

Related People

Takahashi, H: AUTHOR [+2]

Abstract

A record-type change method is described that allows a record type of an existing file to be altered by adding new fields at accessing-time without physical file reorganization. A binary relation file is used to store new field values efficiently. The original record type and this binary relation constitute a virtual record type on which query and update operations are performed.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Record-Type Change Method by Binary Relations

       A record-type change method is described that allows a
record type of an existing file to be altered by adding new fields at
accessing-time without physical file reorganization.  A binary
relation file is used to store new field values efficiently.  The
original record type and this binary relation constitute a virtual
record type on which query and update operations are performed.

      Changing a record type is a prerequisite file operation to
store values of new fields, once the record type is defined.
Conventional physical file reorganization requires data unload/reload
processing, and interactive file access suspends until the
reorganization process is completed.  The problem is the cost of
reorganization which increases in proportion to the number of records
in the file.  Besides, storage use is inefficient when the
reorganized file is sparse where almost all records have null values
for new fields.

      A binary relation file avoids null value fields by storing new
field names with values.  The change of a record type is processed by
record insert/delete operations on the binary relation.  The binary
relation consists of three fields: ID to identify a record in the
original record type, FIELD and VAL to store the name and value,
respectively, of the new field.  The field VAL has three parts: TYPE
to describe the data type such as the integer or character, LENGTH to
specify the byte length of the data, and DATA to store the actual
data.

      Query operations are performed on a virtual record type which
is derived from the original record type and the binary relation at
each accessing-time.  Fig. 1 shows the derivation process.  First,
the transformation process converts a binary relation to a temporal
record type which c...