介绍 #

分布式实时lucene

使用 #

./bin/elasticsearch
curl http://localhost:9200

命令 #

elasticsearch -Ecluster.name=my_cluster_name -Enode.name=my_node_name

概念存储 #

index
    type
        # document中加_type field实现
        # 所以不同type中的field在index要唯一，否则冲突
        # 对field排序会载入所有type的document
        document
            # 对应lucene中的key value倒排文档
            # 对就一个请求的json对象
            field
                mapping
                    # 定义type的field，映射json到document field

设置 #

config/elasticsearch.yml
    action.auto_create_index: -l*, +z*
        # 自动创建，以z开头和非l开头的索引
    action.destructive_requires_name: true
        # 允许通配删index
    http.cors.enables: true
    http.cors.allow-origin: "*"
    cluster.name: c1
    node.name: n1
    node.master: true
    node.data: true
    transport.host: localhost
    transport.tcp.port: 9300
    network.host: 0.0.0.0
        # 修改es监听地址，别的机器也可以访问。同时设置bind_host和publish_host
        # 需要设置transport.host:localhost
    network.bind_host
        # 节点绑定ip
    network.publish_host
        # 发布地址，其它节点通过这个地址通信
    http.port: 9200
    transport.tcp.port
        # 通信端口，默认9300
    discovery.zen.minimum_master_nodes: 2

o-> 可用配置
cluster.name: myES_Cluster
node.name: ESNODE_CYR
node.master: true
node.data: true
transport.host: localhost
transport.tcp.port: 9300
http.port: 9200
network.host: 0.0.0.0
discovery.zen.minimum_master_nodes: 2

接口 #

index #

get/put/delete /index1              # 创建index, get查询，delete删除
    settings
    mappings
    aliases:
put /index1/_mapping/type2
get/put /index1/type2/_mapping      # 创建type或给已有type加mappings
    properties
put /index1/_settings
get/put /index1/type1/1             # 插入doc
    name: "name1"

_cat #

get /_cat/health?v                  # 集群健康
get /_cat/nodes?v                   # 集群节点
get /_cat/indices?v                 # 所有索引

_cluster #

get /_cluster/state?pretty          # 集群状态

_aliases #

post /_aliases                      # 索引别名
    actions:
        add:
            alias: "my_index"
            index: "my_index_v1"
        remove

_template #

put /_template/tpl1
    template: "te*"
        # 匹配所有re开头的index
    settings:
    mappings:

_search #

post /index1/type1/_search
    # from size实时分页
    # scroll快照分页
    ?from=0&size=50
    ?scroll=1m&size=50
        # 过期时间1分钟，每次返回50条
    ?search_type=scan&scroll=1m
        # scroll-scan分页不排序，更快,

_analyze #

post /index1/_analyze
    text: "刘德华"
    analyzer: "analyzer1"

_close #

post /index1/_close
    # 关闭索引，此后可以改settings

_open #

post /index1/_open

_cache #

post /index1/type1/_cache/clear?filter_keys=k1
    # 清空query filter的缓存

数据对象 #

_search #

query
    match
        # 理解如何分词的, 会对field分词再查询
        field1:
            query: "a b"
            operator: "and"
            minimum_should_match: "75%"
                # 匹配的query分词的最低占比
    match_all
        # 默认，会查出所有文档
    multi_match
        query: "a b"
        fields: ["field1", "field2"]
    match_phrase
        # 所有term命中，并且位置邻接
        field1: "a b"
    term
        # 确切查询
        field1: "value1"
    terms
        # 多条件and
        field1: [1,2,3]
    range
        field1:
            gt: 20
            gte:
            lt:
            lte:
    exists:
        field: "field1"
    missing:
        field: "field1"
    regexp
        postcode: "W[0-9].+"
    wildcard
        postcode: "W?F*HW"
    prefix
        # 以某些字符开头
        field1: "a"
    bool
        # 分值计算来自must和should语句, must_not不影响
        must
            match
        must_not
        should: []
        minimum_should_match: 2
    filtered
        query
        filter:
            # filter的field会缓存起来
            ## geo, and, or, not, script, numeric_range的默认不缓存
            term:
                field1: "a"
                _cache_key: "k1"
                _cache: false
            range:
                field1:
                    gte: 0
aggs
    diy1:
        avg:
            field: "field1"
    diy2:
        terms:
            # 聚合查询中的所有term
            field: "field1"
post_filter:
    # 对搜索结果进行过滤
    term:
        field1: "a"
sort: []
    # 默认升序，_score默认降序
    field1
        order: "desc"
            # asc
        mode: "min"
            # 对数组元素排序时的取值, 还有max, sum, avg, median
        missing: "field1"
    "_score",
highlight
    pre_tags: ["<tag1>"]
    post_tags: ["</tag1>"]
    fields:
        content: {}
simple_query_string:
    query: ""
    analyzer:
    fields: ["body^5", "_all"]
    default_operator: "and"

mappings #

type1:
    dynamic: true
        # 默认true,自动给未知field建索引
        # false: 忽略未知field， strict: 未知field报错
    include_in_all: false
        # 默认不include
    _all:
        # meta field
        enabled: false
                # 关闭all作用域
        analyzer:
                # 其实是search_analyzer
        term_vector: no
                # 对field建立词频向量空间
        store: "false"
    _source:
        #  是否保存内容
        enabled: true
    properties:
        field1:
            type: “text”
                # text分词，keyword不分词，numeric, date, string
                # multi_field可定义多个field
            fields:[]
                field1:
                    type
            store: "yes"
            index: "not_analyzed"
                # analyzed
            analyzer: "ik_max_word"
            search_analyzer: "ik_max_word"
                # 默认为analyzer
            include_in_all: "true"
                # 是否加入_all作用域
            boost: 8

aliases #

alias1:
    filter:
        term: user: "kimchy"
    routing: "kimchy"

settings #

# 有些设置不能动态修改
index:
    number_of_shards: 3
    number_of_replicas: 2
    max_result_window: 10000
        # from + size的上限，默认10000
    analysis:
        tokenizer:
            # 处理原始输入
            tokenizer1
                type: "pinyin"
                pinyin_field1:
        filter:
            # tokenizer作为输入
            filter1:
                type: "pinyin"
                pinyin_field1:
        analyzer:
            # 组合tokenizer和filter
            analyzer1:
                type: "custom"
                tokenizer: "ik_smart"
                filter: ["filter1", "word_delimiter"]

插件 #

使用
    复制到/plugins
    场景中，指定type:"xx"使用
## ik
    介绍
        elasticsearch-analysis-ik
    安装
        mvn package
        unzip -d /elasticsearch/plugins/ik ./target/releases/elasticsearch-analysis-ik-1.8.0.zip
        重启elasticsearch
    分词器
        ik_max_word
            curl -XGET 'http://localhost:9200/_analyze?pretty&analyzer=ik_max_word' -d '联想是全球最大的笔记本厂商'
        ik_smart
            curl -XGET 'http://localhost:9200/_analyze?pretty&analyzer=ik_smart' -d '联想是全球最大的笔记本厂商'
    mapping type
        {
        "properties": {
            "content": {
            "type": "text",
            "store": "no",
            "term_vector": "with_positions_offsets",
            "analyzer": "ik_smart",
            "search_analyzer": "ik_smart",
            "include_in_all": "true",
            "boost": 8
            }
        }
        }
## pinyin
    介绍
        elasticsearch-analysis-pinyin


    o->
    "analysis" : {
        "analyzer" : {
            "pinyin_analyzer" : {
                "tokenizer" : "my_pinyin",
                "filter" : "word_delimiter"
            }
        },
        "tokenizer" : {
            "my_pinyin" : {
                # 单字
                "type" : "pinyin",
                "first_letter" : "none",
                "padding_char" : " "
            },
            "my_pinyin_fisrt_letter" : {
                # 首字母, 如北京为bj
                "type" : "pinyin",
                "first_letter" : true,
                "padding_char" : " "
            },
        }
    }
    o-> pinyin
    "analysis" : {
        "tokenizer" : {
            "my_pinyin" : {
                "type" : "pinyin",
                "keep_separate_first_letter" : false,
                "keep_full_pinyin" : true,
                "keep_original" : true,
                "limit_first_letter_length" : 16,
                "lowercase" : true,
                "remove_duplicated_term" : true
            }
        },
        "analyzer" : {
            "pinyin_analyzer" : {
                "tokenizer" : "my_pinyin"
            }
        }
    }
    "properties": {
        "name": {
            "type": "keyword",
            "fields": {
                "pinyin": {
                    "type": "text",
                    "store": "no",
                    "term_vector": "with_offsets",
                    "analyzer": "pinyin_analyzer",
                    "boost": 10
                }
            }
        }
    }

    o-> ik-pinyin
    "analysis": {
        "filter": {
            "pinyin1": {
                "type": "pinyin"
            }
        },
        "analyzer": {
            "ik_pinyin_analyzer": {
                "filter": ["pinyin1","word_delimiter"],
                "type": "custom",
                "tokenizer": "ik_smart"
            }
        }
    },

工具 kopf bigdesk head 使用 https://github.com/mobz/elasticsearch-head cnpm i npm i -g grunt-cli grunt server curl localhost:9100 配置 Gruntfile.js port:9100 client olivere/elastic Search # SearchService Do Index Query Sort From Pretty Index # IndexService Do Index Type Id BodyJson Refresh Suggest # SuggestService query SimpleQueryString