Neoview Control Query Default (CQD) Reference Guide (R2.5)

Setting this CQD to a smaller value means that when NECESSARY keyword or automation

is used, the maximum sample size is smaller. Update statistics may run faster, but may

also create less accurate histograms. Setting this CQD to a larger value results in larger

samples, with update statistics running longer, while providing more accurate histograms.

Impact

SystemLevel

Not applicableConflicts/Synergies

Not applicableAddressing the real

problem

Neoview Release 2.4 SP1Introduced in release

Not applicableDeprecated in release

USTAT_OBSOLETE_PERCENT_ROWCOUNT

Update Statistics and ReorgCategory

Controls the threshold of change after which a table has histograms regenerated when

update statistics automation, or the NECESSARY keyword, is used.

If the total number of rows of a table has changed more than

USTAT_OBSOLETE_PERCENT_ROWCOUNT %, and has histograms that have been read

less than USTAT_MAX_READ_AGE_IN_MIN minutes ago, these histograms are

regenerated. In addition, if the sum of the number of rows that have been updated, deleted,

or inserted exceeds USTAT_OBSOLETE_PERCENT_ROWCOUNT % of the table, the same

occurs.

Description

0 through 100

The default value is ‘15’.

Values

Setting this CQD to a smaller value means that a smaller change to the table causes

histograms to be regenerated when using NECESSARY and automation – and therefore

regeneration of histograms occurs more often. Setting this CQD to a larger value requires

a lot more changes to a table before its histogram is regenerated. It can be used to influence

how often update statistics is run for tables and therefore the total time it takes to update

all statistics.

Usage

Not applicableProduction usage

Depending on this setting and based on the frequency of updates to tables, their histograms

are updated more often or less often. This can impact the time it takes to update histograms

for all the tables on the one hand, but can leave histograms stale for longer and result in

poorer query plans on the other.