Hi Carl, Thanks for your interest on purge functionality in CouchDB.
The _purge endpoint allows operators to specify the document ID and its revision to be purged. In the request, multiple documents can be specified. So it is not necessary to read through the whole database because the documents to be purged can be located directly. Furthermore, they can be purged/permanently deleted from database. It should be mentioned as well that beside document, its secondary index, and other stuff will be also deleted. However, these are aysnc-ed. Regarding the performance or efficiency, it is not worse than the whole database scan. Normally, it just depends on the deletion of documents in database. For more details, you may want to read below. https://docs.couchdb.org/en/main/cluster/purging.html#clustered-purge https://docs.couchdb.org/en/main/api/database/misc.html#post--db-_purge Best, Peng Hui Jiang From: "Carl-Clemens Ebinger" <p...@ebinger.cc> To: dev@couchdb.apache.org Date: 2021/09/14 11:25 PM Subject: [EXTERNAL] Question on Implementation of purge. Hello list, I'm interestested to know how the purge-function works. Is a read-operation of the whole database necessary, and afterwards an rewrite of all data which is not deleted? How efficient is the purge-Operation? Best, Carl-Clemens Ebinger
