The old code assumes that post-aggregator prefixes are one character
long followed by numbers. This isn't always true (we may pad with
underscores to avoid conflicts). Instead, the new code uses a different
base prefix for sort-project postaggregators ("s" instead of "p") and
uses the usual Calcites.findUnusedPrefix function to avoid conflicts.

[ Full content available at: 
https://github.com/apache/incubator-druid/pull/6257 ]
This message was relayed via gitbox.apache.org for [email protected]

Reply via email to