如何在spark executor处收集所有记录并将其作为批处理进行处理 [英]How to collect all records at spark executor and process it as batch