Browse Prior Art Database

Batching Algorithm in Files and Hierarchic Data Bases

IP.com Disclosure Number: IPCOM000046293D
Original Publication Date: 1983-Jun-01
Included in the Prior Art Database: 2005-Feb-07
Document File: 5 page(s) / 108K

Publishing Venue

IBM

Related People

Wilson, GM: AUTHOR

Abstract

A method is described for operating a computing apparatus to batch together requests for access to the same dataset or database and to batch down a hierarchic path. (A request is an independent formal statement of need to access data in a file or hierarchic data base.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 30% of the total text.

Page 1 of 5

Batching Algorithm in Files and Hierarchic Data Bases

A method is described for operating a computing apparatus to batch together requests for access to the same dataset or database and to batch down a hierarchic path. (A request is an independent formal statement of need to access data in a file or hierarchic data base.) The method comprises the steps of (a) batching requests which access the same record or segment stored according to a sequential access method (SAM), or virtual sequential access method (VSAM), or a root segment stored in a data language (DL/I) hierarchy only if the ranges of the requests overlap; and (b) with respect to requests for access within a DL/I hierarchy, batching requests at any particular access level only when all of the following conditions are met: the requests are accessing the same record or segment; the requests are batched at the previous level, if any; the requests have ranges at this level; the ranges overlap; and either at least one request has unbounded range at this level and no qualifiers at all for all previous levels, or all requests have no qualifiers at previous levels.

Batching Requirement

A general-purpose digital computer may be operated under control of an application or control program which issues extract requests against source data. To avoid passing possibly voluminous datasets and data bases many times, and thus improve overall product throughput, there may be a need to batch together extract requests which go against the same source data. According to the method herein described, the computing apparatus is operated to process physical sequential (SAM), VSAM key sequential data sets (KSDS) and entry sequential data sets (ESDS), and DL/I data bases. Extract requests may include statements of conditions which must be met by the source data in order for extraction to take place. Analysis of these statements (called qualifiers) provides information as to whether running these extract requests together as a batch offers a savings.

Access Range and Overlap

The access of a particular record or segment is said to have an access range if a series of consecutive records of segments is normally accessed. This access range is unbounded if all records or segments (within parent for non-root DL/I segments) are accessed; it is bounded if not all records or segments are accessed. Further discussion is by data set or data base type:

1. SAM or VSAM ESDS Because the only way to process these dataset types is to start at the beginning and read consecutive records, a request accessing a dataset of this type automatically has a range. If the dataset does not have a defined sort field, or if the request's qualifiers do not specify the sort field, then the range is unbounded. A bounded range can result from a SORTFIELD less-than-constant qualifier (for ascending sequence; for descending sequence a SORTFIELD qualifier of greater-than-constant should be substituted) inasmuch as the fetching of records...