Let's create a JIRA about creating a general n-ary iterator for collections of chunked arrays so this can be addressed in generality. Such an iterator would want to be able to precompute the number of kernel invocations required so that parallelism can be used in some cases
[ Full content available at: https://github.com/apache/arrow/pull/2461 ] This message was relayed via gitbox.apache.org for [email protected]
