流式计算常见模块用法说明

2021-07-21 02:51:26

说明

Kafka Compositor
{
   "name": "streaming.core.compositor.spark.streaming.source.KafkaStreamingCompositor",
   "params": [{
                 "topics":"your topic",
                 "metadata.broker.list":"brokers",
                 "auto.offset.reset": "smallest|largest"
             }]
}

参数说明：

MockInputStreamCompositor

模拟数据源，主要为了方便测试。

{
        "name": "streaming.core.compositor.spark.streaming.source.MockInputStreamCompositor",
        "params": [{
                      "batch-1":["1","2","3"],
                      "batch-2":["1","2","3"],
                      "batch-3":["1","2","3"],
                      "batch-4":["1","2","3"]
                  }]
}

MockInputStreamFromPathCompositor

模拟数据源，主要为了方便测试。可以接入一个外部文件作为mock数据

{
        "name": "streaming.core.compositor.spark.streaming.source.MockInputStreamFromPathCompositor",
        "params": [{"path":"file:///tmp/test.txt"}]
}

SingleColumnJSONCompositor

把一条日志转化一个单列的json文件。

{
        "name": "streaming.core.compositor.spark.streaming.transformation.SingleColumnJSONCompositor",
        "params": [{
            "name": "a"
          }]
}

params.name 则是列名，方便后续的sql使用。

ScalaMapToJSONCompositor

{
        "name": "streaming.core.compositor.spark.streaming.transformation.ScalaMapToJSONCompositor",
        "params": [{}]
}

可以把scala Map转化为JSon

JavaMapToJSONCompositor

{
        "name": "streaming.core.compositor.spark.streaming.transformation.JavaMapToJSONCompositor",
        "params": [{}]
}

可以把java Map转化为JSon

FlatJSONCompositor

{
        "name": "streaming.core.compositor.spark.streaming.transformation.FlatJSONCompositor",
        "params": [{"a":"$['store']['book'][0]['title']"}]
}

从JSON里抽取字段，映射到新的列名上。主要是对复杂JSON结构进行扁平化。语法参考该库JsonPath

NginxParserCompositor

{
        "name": "streaming.core.compositor.spark.streaming.transformation.NginxParserCompositor",
        "params": [{"time":0,"url":1}]
}

Nginx 日志解析工具，按位置给列进行命名。

SQLCompositor

{
        "name": "streaming.core.compositor.spark.streaming.transformation.SQLCompositor",
        "params": [
          {
            "sql": "select a, \"5\" as b from test",
            "outputTableName": "test2"
          }
        ]

}

SQLESOutputCompositor

将数据存储到ES中

{
        "name":"streaming.core.compositor.spark.streaming.output.SQLESOutputCompositor",
        "params":[
          {
            "es.nodes":"",
            "es.resource":"",
            "es.mapping.include":"",
            "timeFormat":"yyyyMMdd"
          }
        ]
}

SQLPrintOutputCompositor(output)

{
        "name": "streaming.core.compositor.spark.streaming.output.SQLPrintOutputCompositor",
        "params": [{}]
}

把处理结果打印到终端控制台。主要是为了调试使用

JSONTableCompositor

{
        "name": "streaming.core.compositor.spark.streaming.transformation.JSONTableCompositor",
        "params": [{
            "tableName": "test"
          }]
}

把字符串(JSON格式)的数据注册成一张表。 params.tableName可以让你指定表名。

ConsoleOutputCompositor

{
        "name": "streaming.core.compositor.spark.streaming.output.ConsoleOutputCompositor",
        "params": [{ }]
}

控制台打印，非SQL类。

SQLCSVOutputCompositor

{
        "name": "streaming.core.compositor.spark.streaming.output.SQLCSVOutputCompositor",
        "params": [{
  "path":"",
  "mode":""
 }]
}

作为CSV 输出，需要前面是一张表。

SQLParquetOutputCompositor

{
        "name": "streaming.core.compositor.spark.streaming.output.SQLParquetOutputCompositor",
        "params": [{
  "path":"",
  "mode":""
 }]
}

作为parquet 输出，需要前面是一张表。

码农公寓

说明