Relevant ticket: SPARK-47180 <https://issues.apache.org/jira/browse/SPARK-47180>: Migrate CSV parsing off of Univocity
What are you specifically proposing? It’s not clear from your email. Are you suggesting making the CSV parser part of the Spark project? How would you change the library to take advantage of Spark internals? Also, why not fork the library and keep it external, or migrate to something else like FastCSV <https://github.com/osiegmar/FastCSV>, Commons CSV, etc? Nick > On Oct 20, 2026, at 5:35 AM, David <[email protected]> wrote: > > Hello, > > The Spark CSV parsing library is unmaintained. > > https://github.com/uniVocity/univocity-parsers > > Trivial with AI to consume it and tailor it more efficiently with Spark > internals. > > Any interest? > > Thanks. >
