Quality of Statistics As a Function of Sample Size - Teradata Database - Teradata Vantage NewSQL Engine

SQL Data Manipulation Language

Product
Teradata Database
Teradata Vantage NewSQL Engine
Release Number
16.20
Published
March 2019
Language
English (United States)
Last Update
2019-05-03
dita:mapPath
fbo1512081269404.ditamap
dita:ditavalPath
TD_DBS_16_20_Update1.ditaval
dita:id
B035-1146
lifecycle
previous
Product Category
Teradata Vantage™

The larger the sample size used to collect statistics, the more likely the sampled statistics are an accurate representation of population statistics. There is a trade-off for this accuracy: the larger the sample size, the longer it takes to collect the statistics.

You cannot use sampling to collect COLUMN statistics for the partitioning columns of row-partitioned tables and you should not use sampling to collect INDEX statistics for those columns. The collect statistics on the system-derived PARTITION or PARTITION#Ln columns.