PySpark HBase / Phoenix集成

Question

我应该将凤凰数据读入pyspark。

编辑：我正在使用Spark HBase转换器：

这是一段代码：

port="2181"
host="zookeperserver"
keyConv = "org.apache.spark.examples.pythonconverters.ImmutableBytesWritableToStringConverter"
valueConv = "org.apache.spark.examples.pythonconverters.HBaseResultToStringConverter"
cmdata_conf = {"hbase.zookeeper.property.clientPort":port, "hbase.zookeeper.quorum": host, "hbase.mapreduce.inputtable": "camel", "hbase.mapreduce.scan.columns": "data:a"}
sc.newAPIHadoopRDD("org.apache.hadoop.hbase.mapreduce.TableInputFormat","org.apache.hadoop.hbase.io.ImmutableBytesWritable","org.apache.hadoop.hbase.client.Result",keyConverter=keyConv,valueConverter=valueConv,conf=cmdata_conf)

追溯：

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/hdp/2.3.0.0-2557/spark/python/pyspark/context.py", line 547, in newAPIHadoopRDD
    jconf, batchSize)
  File "/usr/hdp/2.3.0.0-2557/spark/python/lib/py4j-0.8.2.1-src.zip/py4j/java_gateway.py", line 538, in __call__
  File "/usr/hdp/2.3.0.0-2557/spark/python/lib/py4j-0.8.2.1-src.zip/py4j/protocol.py", line 300, in get_return_value
py4j.protocol.Py4JJavaError: An error occurred while calling z:org.apache.spark.api.python.PythonRDD.newAPIHadoopRDD.
: java.io.IOException: No table was provided.
    at org.apache.hadoop.hbase.mapreduce.TableInputFormatBase.getSplits(TableInputFormatBase.java:130)

任何帮助将非常感激。

谢谢！ /蒂娜

Answer 1

建议使用spark phoenix插件。请查看有关凤凰火花插件here的详细信息

环境：使用AWS EMR 5.10，PySpark进行测试

以下是步骤

1. 在凤凰https://phoenix.apache.org/language/创建表打开Phoenix shell “/usr/lib/phoenix/bin/sqlline.py” DROP TABLE IF EXISTS TableName; CREATE TABLE TableName（DOMAIN VARCHAR主键）; UPSERT INTO TableName（DOMAIN）VALUES（'foo'）;
2. 从https://mvnrepository.com/artifact/org.apache.phoenix/phoenix-core/4.11.0-HBase-1.3下载spark phoenix plugin jar下载spark phoenix plugin jar你需要phoenix - HBase - client.jar，我根据我的phoenix和hbase版本使用了phoenix-4.11.0-HBase-1.3-client.jar
3. 从您的hadoop主目录，设置以下变量： phoenix_jars = /家庭/用户/ Apache的凤凰4.11.0-HBase的-1.3斌/凤凰-4.11.0-HBase的-1.3-client.jar中
4. 启动PySpark shell并在Driver和executor类路径中添加依赖项 pyspark --jars $ {phoenix jars} --conf spark.executor.extraClassPath = $ {phoenix jars}

- 创建ZooKeeper URL，替换为您的集群zookeeper仲裁，您可以从hbase-site.xml进行检查

emrMaster = "ZooKeeper URL" 

df = sqlContext.read \
.format("org.apache.phoenix.spark") \
.option("table", "TableName") \
.option("zkUrl", emrMaster) \
.load() 

df.show()
df.columns
df.printSchema()
df1=df.replace(['foo'], ['foo1'], 'DOMAIN')
df1.show() 

df1.write \
  .format("org.apache.phoenix.spark") \
  .mode("overwrite") \
  .option("table", "TableName") \
  .option("zkUrl", emrMaster) \
  .save()

Answer 2

有两种方法可以做到这一点 ： 1）由于Phoenix有一个JDBC层，你可以使用Spark JdbcRDD从Spark https://spark.apache.org/docs/1.3.0/api/java/org/apache/spark/rdd/JdbcRDD.html中的Phoenix读取数据

2）使用Spark HBAse转换器：https://github.com/apache/spark/blob/master/examples/src/main/scala/org/apache/spark/examples/pythonconverters/HBaseConverters.scala https://github.com/apache/spark/tree/master/examples/src/main/python