default_statistics_target

default_statistics_target is the PostgreSQL parameter that controls the granularity of the statistics that ANALYZE builds for each column. The default value is 100.

How it works #

ANALYZE builds two statistical structures per column, used by the optimizer:

• Most common values (MCV): the list of the most frequent values, with their respective frequencies
• Histogram: the distribution of the remaining values, divided into equal-population buckets

default_statistics_target determines how many elements these structures can hold. With the value 100: up to 100 values in the MCV list and up to 100 buckets in the histogram.

The number of sampled rows is a separate matter and depends on the target: it’s roughly 300 × default_statistics_target. With the default of 100, ANALYZE reads ~30,000 rows per column; with a target of 500, ~150,000. So raising the target increases both the granularity of the statistics and the cost of ANALYZE.

When to increase it #

For small tables or tables with uniform distribution, 100 samples are sufficient. For large tables with skewed distribution — where a few values dominate most rows — 100 samples can give a distorted picture, leading the optimizer to wrong cardinality estimates.

You can increase the target at the column level:

ALTER TABLE orders ALTER COLUMN status SET STATISTICS 500;
ANALYZE orders;

Values between 500 and 1000 significantly improve estimate quality on columns with non-uniform distribution.

Practical limits #

Beyond 1000 the benefit is marginal and ANALYZE itself becomes slower, because it needs to sample more rows and build larger structures. It’s a fine-tuning adjustment: apply it only to columns that actually cause wrong estimates, not to every column in every table.