I Did some tweaking to ExplicitColumnsFilter (former
ExplicitScanReplacementFilter) and now  getting 1.45M rows per sec ( > 350%
speed up compared to default implementation). The raw scan (no columns
specified) runs at 1.6M rows per sec.  Its only 10% performance hit.
Phoenix team should be happy today. I am opening JIRA.

Reply via email to