更新时间:2024-06-27 GMT+08:00
分享

使用Spark执行Hudi样例程序(Scala)

下面代码片段仅为演示,具体代码参见:com.huawei.bigdata.hudi.examples.HoodieDataSourceExample。

插入数据:

def insertData(spark: SparkSession, tablePath: String, tableName: String, dataGen: HoodieExampleDataGenerator[HoodieAvroPayload]): Unit = {
val commitTime: String = System.currentTimeMillis().toString
val inserts = dataGen.convertToStringList(dataGen.generateInserts(commitTime, 20))  
spark.sparkContext.parallelize(inserts, 2)
val df = spark.read.json(spark.sparkContext.parallelize(inserts, 1))df.write.format("org.apache.hudi").
      options(getQuickstartWriteConfigs).
      option(PRECOMBINE_FIELD_OPT_KEY, "ts").
      option(RECORDKEY_FIELD_OPT_KEY, "uuid").
      option(PARTITIONPATH_FIELD_OPT_KEY, "partitionpath").
      option(TABLE_NAME, tableName).
      mode(Overwrite).
      save(tablePath)}

查询数据:

def queryData(spark: SparkSession, tablePath: String, tableName: String, dataGen: HoodieExampleDataGenerator[HoodieAvroPayload]): Unit = {
val roViewDF = spark.
      read.
      format("org.apache.hudi").
      load(tablePath + "/*/*/*/*")
roViewDF.createOrReplaceTempView("hudi_ro_table")
spark.sql("select fare, begin_lon, begin_lat, ts from  hudi_ro_table where fare > 20.0").show()
//  +-----------------+-------------------+-------------------+---+  
//  |             fare|          begin_lon|          begin_lat| ts|  
//  +-----------------+-------------------+-------------------+---+  
//  |98.88075495133515|0.39556048623031603|0.17851135255091155|0.0|  
//  ...  
spark.sql("select _hoodie_commit_time, _hoodie_record_key, _hoodie_partition_path, rider, driver, fare from  hudi_ro_table").show()
//  +-------------------+--------------------+----------------------+-------------------+--------------------+------------------+  
//  |_hoodie_commit_time|  _hoodie_record_key|_hoodie_partition_path|              rider|              driver|              fare|  
//  +-------------------+--------------------+----------------------+-------------------+--------------------+------------------+  
//  |     20191231181501|31cafb9f-0196-4b1...|            2020/01/02|rider-1577787297889|driver-1577787297889| 98.88075495133515|  
//  ...
}

更新数据:

def updateData(spark: SparkSession, tablePath: String, tableName: String, dataGen: HoodieExampleDataGenerator[HoodieAvroPayload]): Unit = {
val commitTime: String = System.currentTimeMillis().toString
val updates = dataGen.convertToStringList(dataGen.generateUpdates(commitTime, 10))
val df = spark.read.json(spark.sparkContext.parallelize(updates, 1))
df.write.format("org.apache.hudi").
      options(getQuickstartWriteConfigs).
      option(PRECOMBINE_FIELD_OPT_KEY, "ts").
      option(RECORDKEY_FIELD_OPT_KEY, "uuid").
      option(PARTITIONPATH_FIELD_OPT_KEY, "partitionpath").
      option(TABLE_NAME, tableName).
      mode(Append).
      save(tablePath)}

增量查询:

def incrementalQuery(spark: SparkSession, tablePath: String, tableName: String) {
import spark.implicits._
val commits = spark.sql("select distinct(_hoodie_commit_time) as commitTime from hudi_ro_table order by commitTime").map(k => k.getString(0)).take(50)
val beginTime = commits(commits.length - 2) 

val incViewDF = spark.
      read.
      format("org.apache.hudi").
      option(QUERY_TYPE_OPT_KEY, QUERY_TYPE_INCREMENTAL_OPT_VAL).
      option(BEGIN_INSTANTTIME_OPT_KEY, beginTime).
      load(tablePath)
incViewDF.createOrReplaceTempView("hudi_incr_table")
spark.sql("select `_hoodie_commit_time`, fare, begin_lon, begin_lat, ts from hudi_incr_table where fare > 20.0").show()}

特定时间点查询:

def pointInTimeQuery(spark: SparkSession, tablePath: String, tableName: String) {
import spark.implicits._
val commits = spark.sql("select distinct(_hoodie_commit_time) as commitTime from  hudi_ro_table order by commitTime").map(k => k.getString(0)).take(50)
val beginTime = "000" 
// Represents all commits > this time.  
val endTime = commits(commits.length - 2) 
// commit time we are interested in  
//incrementally query data  
val incViewDF = spark.read.format("org.apache.hudi").
      option(QUERY_TYPE_OPT_KEY, QUERY_TYPE_INCREMENTAL_OPT_VAL).
      option(BEGIN_INSTANTTIME_OPT_KEY, beginTime).
      option(END_INSTANTTIME_OPT_KEY, endTime).
      load(tablePath)
incViewDF.createOrReplaceTempView("hudi_incr_table")  
spark.sql("select `_hoodie_commit_time`, fare, begin_lon, begin_lat, ts from  hudi_incr_table where fare > 20.0").show()}

相关文档