总结 :
spark–submit 提交的优先级 < scala/java代码中的配置参数 < spark SQL hint
spark submit 中提交参数
#!/usr/bin/env bash
source /home/work/batch_job/product/common/common.sh
spark_version="/home/work/opt/spark"
export SPARK_CONF_DIR=${spark_version}/conf/
spark_shell="/home/opt/spark/spark3-client/bin/spark-shell"
spark_sql="/home/work/opt/spark/spark3-client/bin/spark-sql"
echo ${spark_sql}
echo ${spark_shell}
${spark_shell} --master yarn
--queue test
--name "evelopment_sun-data-new_spark_shell"
--conf "spark.speculation=true"
--conf "spark.network.timeout=400s"
--conf "spark.executor.cores=2"
--conf "spark.executor.memory=4g"
--conf "spark.executor.instances=300"
--conf "spark.driver.maxResultSize=4g"
--conf "spark.sql.shuffle.partitions=800"
--conf "spark.driver.extraJavaOptions=-Dfile.encoding=utf-8"
--conf "spark.executor.extraJavaOptions=-Dfile.encoding=utf-8"
--conf "spark.driver.memory=8g"
--conf "spark.sql.autoBroadcastJoinThreshold=-1"
--conf "spark.sql.turing.pooledHiveClientEnable=false"
--conf "spark.sql.hive.metastore.jars=/home/work/opt/spark/spark3-client/hive_compatibility/*"
--conf "spark.driver.extraClassPath=./__spark_libs__/hive-extensions-2.0.0.0-SNAPSHOT.jar:./hive_jar/parquet-hadoop-bundle-1.6.0.jar:/home/work/opt/spark/spark3-client/hive_compatibility/parquet-hadoop-bundle-1.6.0.jar"
--conf spark.hadoop.mapreduce.fileoutputcommitter.algorithm.version=2
--conf "spark.sql.legacy.timeParserPolicy=LEGACY"
--conf "spark.sql.storeAssignmentPolicy=LEGACY"
--conf spark.executor.extraJavaOptions="-XX:+UseG1GC"
--jars ./online-spark-1.0-SNAPSHOT.jar
scala/java代码中的配置参数
val conf = new SparkConf().setAppName(s"production_data-new_UserOverview_${event_day}")
val spark = SparkSession.builder().config("spark.debug.maxToStringFields", "500").config(conf).getOrCreate()
SQL hint
SELECT /*+ MERGEJOIN(t2) */ * FROM t1 INNER JOIN t2 ON t1.key = t2.key;
Hints – Spark 3.5.0 hints Documentation
原文地址:https://blog.csdn.net/u010003835/article/details/134666922
本文来自互联网用户投稿,该文观点仅代表作者本人,不代表本站立场。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。
如若转载,请注明出处:http://www.7code.cn/show_31990.html
如若内容造成侵权/违法违规/事实不符,请联系代码007邮箱:suwngjj01@126.com进行投诉反馈,一经查实,立即删除!
声明:本站所有文章,如无特殊说明或标注,均为本站原创发布。任何个人或组织,在未征得本站同意时,禁止复制、盗用、采集、发布本站内容到任何网站、书籍等各类媒体平台。如若本站内容侵犯了原著者的合法权益,可联系我们进行处理。