PostgreSQL version:

replacement_sort_tuples

Deprecated

When the number of tuples to be sorted is smaller than this number, a sort will produce its first output run using replacement selection rather than quicksort. This may be useful in memory-constrained environments where tuples that are input into larger sort operations have a strong physical-to-logical correlation. Note that this does not include input tuples with an inverse correlation. It is possible for the replacement selection algorithm to generate one long run that requires no merging, where use of the default strategy would result in many runs that must be merged to produce a final sorted output. This may allow sort operations to complete sooner.

The default is 150,000 tuples. Note that higher values are typically not much more effective, and may be counter-productive, since the priority queue is sensitive to the size of available CPU cache, whereas the default strategy sorts runs using a cache oblivious algorithm. This property allows the default sort strategy to automatically and transparently make effective use of available CPU cache.

Setting maintenance_work_mem to its default value usually prevents utility command external sorts (e.g., sorts used by CREATE INDEX to build B-Tree indexes) from ever using replacement selection sort, unless the input tuples are quite wide.

Recommendations

Disable, this setting will be removed from Postgres 11.

At postgresqlCO.NF (OnGres) we value your privacy and treat all data very seriously. We're fully GDPR compliant, and we continuously monitor and improve our data storage, retention and compliance mechanisms.

This web page does not, however, store any PII (Personally Identifiable Information). The only service that stores any data is Google Analytics, and we use it to gather analytics of the web page.

This website contains some data from the official documentation of the PostgreSQL.org project, and from Annotated.Conf, used with permission.

If you have any question or concern about our terms of service or privacy policy, please contact us at dataprotection _at_ ongres _dot_ com.

OK