执行spark-sql

spark-sql --master yarn --deploy-mode client \
  --executor-cores 8 \
  --num-executors 70 \
  --executor-memory 16G \
  --driver-memory 8G \
  --conf spark.executor.memoryOverhead=4G \
  --conf spark.debug.maxToStringFields=100 \
  --conf spark.driver.maxResultSize=10G

SELECT t1.creative_id as creative_id, t1.pos_id, imp, click, conversion, expect_income, income FROM (SELECT ad_creative_id as creative_id, pos_id, sum(ad_exp_cnt) as imp, sum(ad_clk_cnt) as click, SUM(ad_consume)/100000 as income FROM dm_cc_dsp_user_action_inc_d WHERE dt>= ‘2023-01-31’ GROUP BY ad_creative_id,pos_id having sum(ad_exp_cnt)>0 )t1 LEFT JOIN (SELECT ad_creative_id as creative_id, pos_id, sum(ad_target_cost)/100 as expect_income, SUM (ad_oCPX_resp_cnt) as conversion FROM dm_cc_dsp_user_action_transform_inc_d WHERE dt>= ‘2023-01-31’ and ad_campaign_id is not null GROUP BY ad_creative_id, pos_id )AS t2 ON t1.creative_id=t2.creative_id and t1.pos_id =t2.pos_id