Eel初次使用示例(使用Scala读取Hive表)

• 前言
• 示例
• • Local
  • • CSV File
  • Hive
  • • Parquet
    • ORC
• 后记

最近项目需要用Scala读取Hive的Table，但是Spark和Hive的API比较重量级，性能不够。于是调研，发现了一个更加底层的框架Eel：

为了便于理解，先从读取简单的CSV开始

我自己新建了一个CSV文件

school_number,student_name
1,Bernoulli
2,Newton
3,Leibniz

sbt导入依赖libraryDependencies += "io.eels" % "eel-core_2.12" % "1.2.4"后，读取的代码

import io.eels.{DataTypeRule, SchemaInferrer}
import io.eels.component.csv.CsvSource
import io.eels.schema.{IntType, StringType}

import java.nio.file.Paths

object CsvReaderDemo {
 def main(args: Array[String]): Unit = {
    val csvFilePath = Paths.get("./datas/students.csv")
    // 打印Schema，默认都是Stirng类型
    CsvSource(csvFilePath).toDataStream()
      .schema.fields.foreach(f => println(f))
    // 手动指定字段和类型之间的映射
    val filedTypeMapping = Seq(
      DataTypeRule("school_number", IntType.Signed),
      DataTypeRule("student_name", StringType))
    val schemaInferrer = SchemaInferrer(
      StringType, // 默认类型
      filedTypeMapping)
    // 打印数据
    CsvSource(csvFilePath).withSchemaInferrer(schemaInferrer)
      .toDataStream()
      .filter(_.values(0).toString.toInt > 0) // 根据索引
      .filter(_.get("student_name").equals("Leibniz")) // 根据字段名判断
      //      .filter("student_name", _ == "Leibniz") // 也可以写成这样
      .collect
      .foreach(row => println(row))
  }
}

打印的结果如下

需要创建HMS的客户端

val hiveConf = new HiveConf()
hiveConf.set(ConfVars.metaSTOREURIS.varname, "thrift://localhost:9083")
implicit val hivemetaStoreClient = new HivemetaStoreClient(hiveConf)

准备数据

-- 建表
CREATE EXTERNAL TABLE IF NOT EXISTS `eel_test.person` (
   `NAME` string,
   `AGE` int,
   `SALARY` decimal(38,5),
   `CREATION_TIME` timestamp)
PARTITIonED BY (`title` string)
ROW FORMAT SERDE
   'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'
STORED AS INPUTFORMAT
   'org.apache.hadoop.hive.ql.io.parquet.MapredParquetInputFormat'
OUTPUTFORMAT
   'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat'
LOCATION '/client/eel_test/persons';
 -- 插入数据
INSERT INTO eel_test.person PARTITION (title='Mr') SELECT 'Fred', 50, 50000.99, CURRENT_TIMESTAMP();
INSERT INTO eel_test.person PARTITION (title='Mr') SELECT 'Gary', 50, 20000.34, CURRENT_TIMESTAMP();
INSERT INTO eel_test.person PARTITION (title='Mr') SELECT 'Alice', 50, 99999.98, CURRENT_TIMESTAMP()

读取数据的代码

HiveSource("eel_test", "eel_test")
  .toDataStream()
  .collect
  .foreach(row => println(row))

准备数据

-- 建表
CREATE EXTERNAL TABLE IF NOT EXISTS eel_test.test_orc(
	id int,
	name string
)
PARTITIonED BY (create_date string)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ','
STORED AS ORC
LOCATION '/client/eel_test/test_orc'
 -- 插入数据
INSERT INTO eel_test.test_orc PARTITION(create_date='20210101') 
SELECT  1, 'Alice' 
UNIOn 
SELECT 2, 'Fred' 
UNIOn 
SELECT 3, 'Gary';

读取数据的代码

HiveSource("eel_test", "eel_test")
  .toDataStream()
  .collect
  .foreach(row => println(row))

Eel还能直接读取HDFS，更多例子可以参考github