Read Ahead Count - Advanced SQL Engine - Teradata Database

Database Utilities

Product
Advanced SQL Engine
Teradata Database
Release Number
17.05
Published
January 2021
Language
English (United States)
Last Update
2021-01-23
dita:mapPath
ynh1604715438919.ditamap
dita:ditavalPath
hoy1596145193032.ditaval
dita:id
B035-1102
lifecycle
previous
Product Category
Teradata Vantageā„¢

Purpose

Specifies the number of data blocks that will be preloaded into memory in advance of the current data block when performing sequential table scans if the ReadAhead field is set to TRUE.

Field Group

Performance

Valid Range

1 through 100 data blocks

Default

1 data block

Changes Take Effect

After the DBS Control Record has been written.

Usage Notes

Increasing the Read Ahead Count field setting can reduce the amount of CPU time spent waiting for read operations to finish. Read Ahead Count should be set high enough to make the typical sequential scan limited only by the CPU, rather than by the read I/O.

The CPU must work harder when data blocks are large or row selection criteria are complex. Consequently, if either of these conditions exist, read ahead counts can be lower.

For example, if a large default data block size is used, most of the tables will consist of large data blocks, and the default ReadAheadCount will suffice. If the default data block size is made very small, most of the tables will consist of small data blocks and system performance might benefit by increasing the ReadAheadCount to 25 or higher.