一，测试pipeline，按逗号拆分字符串为数组，然后对数组的每个项去空格。

POST _ingest/pipeline/_simulate
{
  "pipeline" :
  {
    "description": "_description",
    "processors": [
      {
        "set" : {
          "field" : "field2",
          "value" : "_value"
        }
      },
      {
        "split": {
          "field": "words",
          "separator": "," 
        }
      },
      {
        "foreach" : {
          "field" : "words",
          "processor" : {
            "trim": {
              "field" : "_ingest._value"
            }
          }
        }
      }
    ]
  },
  "docs": [
    {
      "_index": "index",
      "_id": "id",
      "_source": {
        "foo": " bar ",
        "words":"hello , world , hello2  hadoop "
      }
    },
    {
      "_index": "index",
      "_id": "id",
      "_source": {
        "foo": "rab",
        "words":"hello, world , hello2 , flink "
      }
    }
  ]
}

二，设置pipeline

PUT _ingest/pipeline/split_and_trim
{
  "description" : "describe pipeline",
  "processors": [
      {
        "set" : {
          "field" : "field2",
          "value" : "_value"
        }
      },
      {
        "split": {
          "field": "words",
          "separator": "," 
        }
      },
      {
        "foreach" : {
          "field" : "words",
          "processor" : {
            "trim": {
              "field" : "_ingest._value"
            }
          }
        }
      }
    ]
}

三，创建index时设置pipeline

PUT twitter
{
    "settings" : {
        "index" : {
            "number_of_shards" : 3, 
            "number_of_replicas" : 2 ,
            "default_pipeline": "split_and_trim"
        }
    }
}

四，更新时指定pipeline

POST twitter2/_update_by_query?pipeline=split_and_trim

五，索引文档时指定pipeline

PUT twitter2/_doc/2?pipeline=split_and_trim
{
  "words":"hadoop222222, good , flink , spark "
}