Found while working on #10799, suggested by @RaduBerinde: when GROUP BY is specified using a column ordinal (e.g. GROUP BY 1) we duplicate the rendered expression in the select node instead of reusing the existing render. This could be perhaps optimized.
Found while working on #10799, suggested by @RaduBerinde: when GROUP BY is specified using a column ordinal (e.g.
GROUP BY 1) we duplicate the rendered expression in the select node instead of reusing the existing render. This could be perhaps optimized.