spark submit读写hudi

yiduwangkai

发布于 2021-09-10 10:58:32

1.1K0

发布于 2021-09-10 10:58:32

文章被收录于专栏：大数据进阶大数据进阶

测试数据：hudi官方自带的batch_1.json

环境准备：

mysql
kafka：kafka_2.13-2.7.0
hadoop：hadoop-2.10.1
hive：hive-2.3.8 4.1.metastore 4.2hiveserver2

导入工具kafkacat

数据导入：cat batch_1.json | kafkacat -b localhost:9092 -t stock_ticks -P

topic查看：kafkacat -L -b localhost:9092 -t stock_tick

元数据查看：kafkacat -b localhost:9092 -L -J | jq

schema准备：hudi官方自带的schema.avsc

spark这里我们用的是spark-2.4.8-bin-hadoop2.7

执行命令：

1. 非自动同步

bin/spark-submit \
--master yarn \
--class org.apache.hudi.utilities.deltastreamer.HoodieDeltaStreamer /Users/wangkai/apps/install/hudi/hudi-utilities-bundle_2.11-0.9.0-SNAPSHOT.jar \
--table-type COPY_ON_WRITE \
--source-class org.apache.hudi.utilities.sources.JsonKafkaSource \
--source-ordering-field ts \
--target-base-path hdfs:///user/hive/warehouse/stock_ticks_cow \
--target-table stock_ticks_cow \
--props file:///Users/wangkai/apps/install/hudi/kafka-source_1.properties \
--schemaprovider-class org.apache.hudi.utilities.schema.FilebasedSchemaProvider

上述命令执行完，我们分别去hdfs和hive中进行查看

hive中空也

执行同步脚本

./run_sync_tool.sh \
--jdbc-url "jdbc:hive2://localhost:10000" \
--user "xxxx" \
--pass "xxxx" \
--partitioned-by dt \
--base-path hdfs:///user/hive/warehouse/stock_ticks_cow \
--database hudi_stock \
--table stock_ticks_cow

执行完去hive中查看

2.自动同步

bin/spark-submit \
--master yarn \
--class org.apache.hudi.utilities.deltastreamer.HoodieDeltaStreamer /Users/wangkai/apps/install/hudi/hudi-utilities-bundle_2.11-0.9.0-SNAPSHOT.jar \
--table-type COPY_ON_WRITE \
--source-class org.apache.hudi.utilities.sources.JsonKafkaSource \
--source-ordering-field ts \
--target-base-path hdfs:///user/hive/warehouse/stock_ticks_cow \
--target-table stock_ticks_cow \
--props file:///Users/wangkai/apps/install/hudi/kafka-source_1.properties \
--schemaprovider-class org.apache.hudi.utilities.schema.FilebasedSchemaProvider
--disable-compaction --enable-hive-sync

这里主要就最后一句：--enable-hive-sync

执行完上述命令hive中就能看到期望中的表

上诉都是针对的copy on write

下面我们同样的步骤描述一下merge on read

1.非自动同步

bin/spark-submit \
--master yarn \
--class org.apache.hudi.utilities.deltastreamer.HoodieDeltaStreamer /Users/wangkai/apps/install/hudi/hudi-utilities-bundle_2.11-0.9.0-SNAPSHOT.jar \
--table-type MERGE_ON_READ \
--source-class org.apache.hudi.utilities.sources.JsonKafkaSource \
--source-ordering-field ts \
--target-base-path hdfs:///user/hive/warehouse/stock_ticks_test_a \
--target-table stock_ticks_test_a \
--props file:///Users/wangkai/apps/install/hudi/kafka-source_1.properties \
--schemaprovider-class org.apache.hudi.utilities.schema.FilebasedSchemaProvider \

同步脚步

./run_sync_tool.sh \
--jdbc-url "jdbc:hive2://localhost:10000" \
--user "xxx" \
--pass "xxx" \
--partitioned-by dt \
--base-path hdfs:///user/hive/warehouse/stock_ticks_mor \
--database hudi_stock \
--table stock_ticks_mor

hive中查看：

2.自动同步

bin/spark-submit \
--master yarn \
--class org.apache.hudi.utilities.deltastreamer.HoodieDeltaStreamer /Users/wangkai/apps/install/hudi/hudi-utilities-bundle_2.11-0.9.0-SNAPSHOT.jar \
--table-type MERGE_ON_READ \
--source-class org.apache.hudi.utilities.sources.JsonKafkaSource \
--source-ordering-field ts \
--target-base-path hdfs:///user/hive/warehouse/stock_ticks_test_a \
--target-table stock_ticks_test_a \
--props file:///Users/wangkai/apps/install/hudi/kafka-source_1.properties \
--schemaprovider-class org.apache.hudi.utilities.schema.FilebasedSchemaProvider \
--disable-compaction --enable-hive-sync

本文参与腾讯云自媒体分享计划，分享自作者个人站点/博客。

如有侵权请联系 cloudcommunity@tencent.com 删除

hive