1，ES的存储结构了解

在ES中，存储结构主要有四种，与传统的关系型数据库对比如下：
index（Indices）相当于一个database
type相当于一个table
document相当于一个row
properties（Fields）相当于一个column

Relational DB -> Databases -> Tables -> Rows -> Columns
Elasticsearch -> Indices -> Types -> Documents -> Fields

2，ES写入测试

写入一个文档（一条数据）

PUT http://192.168.1.32:9200/twitter/tweet/377827236
{
"tweet_id": "555555555555555555555666",
"user_screen_name": "kanazawa_mj",
"tweet": "blog3444444",
"user_id": "377827236",
"id": 214019
}

我们看到path:/twitter/tweet/377827236包含三部分信息：

3，ES查询测试

查询一个文档，包含love，返回50条数据，采用展开的json格式

GET http://192.168.1.32:9200/twitter/tweet/_search?q=tweet:love&size=50&pretty=true
{
  "took" : 20,
  "timed_out" : false,
  "_shards" : {
    "total" : 5,
    "successful" : 5,
    "failed" : 0
  },
  "hits" : {
    "total" : 11639,
    "max_score" : 8.448289,
    "hits" : [
      {
        "_index" : "twitter",
        "_type" : "tweet",
        "_id" : "AV0fnFOX6PBTXc6mRjpL",
        "_score" : 8.448289,
        "_source" : {
          "tweet_id" : "843105177913757697",
          "user_screen_name" : "jessicapalapal",
          "tweet" : "Love, love, love ",
          "user_id" : "740434015",
          "id" : 474551
        }
      },
      {
        "_index" : "twitter",
        "_type" : "tweet",
        "_id" : "AV0fni__6PBTXc6mSeyR",
        "_score" : 8.436986,
        "_source" : {
          "tweet_id" : "695096306763583488",
          "user_screen_name" : "SampsonMariel",
          "tweet" : "Love love love^_^ #ALDUB29thWeeksary",
          "user_id" : "2483556636",
          "id" : 723297
        }
      },
      {
        "_index" : "twitter",
        "_type" : "tweet",
        "_id" : "AV0fmxvV6PBTXc6mQ8Mb",
        "_score" : 8.425938,
        "_source" : {
          "tweet_id" : "835676311637086209",
          "user_screen_name" : "thedaveywavey",
          "tweet" : "Love is love is love is love. ",
          "user_id" : "17191297",
          "id" : 311967
        }
      }
    ]
  }
}

4，ES批量写入测试

• 写入程序，编写Python脚本，生产者和消费者模式，从Mysql数据库读取数据，1000条数据写入一次ES
• 本机环境，Windows，内存占用100M，CPU占用15%
• ES服务，Ubuntu14.04，CPU占用5%，内存较少
• 单进程，5个写入线程，100万行数据，500秒
• 单进程，20个写入线程，100万行数据，500秒
• 补充：据说，修改ES配置，先关闭数据索引，可以提高数据写入速度，尚未测试

5，下一步计划

• ES数据分片机制、搜索参数配置（mapping、filter）等，尚需要根据项目需求，深入学习和测试。
• ES支持的额外功能，例如时间范围搜索、中文简繁体、拼音搜索、GIS位置搜索、英文时态支持等。

6，参考资料

ES的存储结构介绍
https://es.xiaoleilu.com/010_Intro/25_Tutorial_Indexing.html
python操作Elasticsearch
http://www.cnblogs.com/yxpblog/p/5141738.html
Elasticsearch权威指南 - 检索文档
https://es.xiaoleilu.com/010_Intro/30_Tutorial_Search.html

7，附件（Python写入ES代码）

# coding=utf-8
from elasticsearch import Elasticsearch
from elasticsearch.helpers import bulk
import time
import argparse
import sys
reload(sys)
sys.setdefaultencoding('utf-8')

# ES索引和Type名称
INDEX_NAME = "twitter"
TYPE_NAME = "tweet"

# ES操作工具类
class es_tool():
    # 类初始化函数
    def __init__(self, hosts, timeout):
        self.es = Elasticsearch(hosts, timeout=5000)
        pass

    # 将数据存储到es中
    def set_data(self, fields_data=[], index_name=INDEX_NAME, doc_type_name=TYPE_NAME):
        # 创建ACTIONS
        ACTIONS = []
        # print "es set_data length",len(fields_data)
        for fields in fields_data:
            # print "fields", fields
            # print fields[1]
            action = {
                "_index": index_name,
                "_type": doc_type_name,
                "_source": {
                    "id": fields[0],
                    "tweet_id": fields[1],
                    "user_id": fields[2],
                    "user_screen_name": fields[3],
                    "tweet": fields[4]
                }
            }
            ACTIONS.append(action)

        # print "len ACTIONS", len(ACTIONS)
        # 批量处理
        success, _ = bulk(self.es, ACTIONS, index=index_name, raise_on_error=True)
        print('Performed %d actions' % success)

# 读取参数
def read_args():
    parser = argparse.ArgumentParser(description="Search Elastic Engine")
    parser.add_argument("-i", dest="input_file", action="store", help="input file1", required=False, default="./data.txt")
    # parser.add_argument("-o", dest="output_file", action="store", help="output file", required=True)
    return parser.parse_args()

# 初始化es，设置mapping
def init_es(hosts=[], timeout=5000, index_name=INDEX_NAME, doc_type_name=TYPE_NAME):
    es = Elasticsearch(hosts, timeout=5000)
    my_mapping = {
        TYPE_NAME: {
            "properties": {
                "id": {
                    "type": "string"
                },
                "tweet_id": {
                    "type": "string"
                },
                "user_id": {
                    "type": "string"
                },
                "user_screen_name": {
                    "type": "string"
                },
                "tweet": {
                    "type": "string"
                }
            }
        }
    }
    try:
        # 先销毁，后创建Index和mapping
        delete_index = es.indices.delete(index=index_name)  # {u'acknowledged': True}
        create_index = es.indices.create(index=index_name)  # {u'acknowledged': True}
        mapping_index = es.indices.put_mapping(index=index_name, doc_type=doc_type_name,
                                                    body=my_mapping)  # {u'acknowledged': True}
        if delete_index["acknowledged"] != True or create_index["acknowledged"] != True or mapping_index["acknowledged"] != True:
            print "Index creation failed..."
    except Exception, e:
        print "set_mapping except", e

# 主函数
if __name__ == '__main__':
    # args = read_args()
    # 初始化es环境
    init_es(hosts=["192.168.1.32:9200"], timeout=5000)
    # 创建es类
    es = es_tool(hosts=["192.168.1.32:9200"], timeout=5000)
    # 执行写入操作
    tweet_list = [("111","222","333","444","555"), ("11","22","33","44","55")]
    es.set_data(tweet_list)