最初接觸Elasticsearch是在ELK日志系統(tǒng)的建設(shè)中,隨著對(duì)日志數(shù)據(jù)的消費(fèi)越來(lái)越多,被其強(qiáng)大的數(shù)據(jù)搜索和分析能力所吸引;后來(lái),在用戶(hù)行為數(shù)據(jù)采集系統(tǒng)中,使用Elasticsearch做核心數(shù)據(jù)存儲(chǔ)和實(shí)時(shí)聚合分析;再后來(lái),使用Elasticsearch搭建了產(chǎn)品的搜索服務(wù)。目前來(lái)看,Elasticsearch在這三個(gè)系統(tǒng)中表現(xiàn)都很靈活和優(yōu)異,沒(méi)有讓我們失望,而在系統(tǒng)建設(shè)中,我們也遇到過(guò)不少問(wèn)題,有基本概念的迷惑、操作方法、部署、性能等等各個(gè)方面。本文著重對(duì)Elasticsearch在應(yīng)用層面上的使用進(jìn)行總結(jié),搞清楚WHAT和HOW兩個(gè)層面,即是什么、怎么用。
NOTE:本文所述的概念和方法均在Elasticsearch2.3版本下。
基本概念
Elasticsearch is a distributed, RESTful search and analytics engine capable of solving a growing number of use cases. As the heart of the Elastic Stack, it centrally stores your data so you can discover the expected and uncover the unexpected.
這是官方對(duì)Elasticsearch的定位。通俗的講,Elasticsearch就是一款面向文檔的NoSQL數(shù)據(jù)庫(kù),使用JSON作為文檔序列化格式。但是,它的高級(jí)之處在于,使用Lucene作為核心來(lái)實(shí)現(xiàn)所有索引和搜索的功能,使得每個(gè)文檔的內(nèi)容都可以被索引、搜索、排序、過(guò)濾。同時(shí),提供了豐富的聚合功能,可以對(duì)數(shù)據(jù)進(jìn)行多維度分析。對(duì)外統(tǒng)一使用REST API接口進(jìn)行溝通,即Client與Server之間使用HTTP協(xié)議通信。
首先,來(lái)看看在存儲(chǔ)上的基本概念,這里將其與MySQL進(jìn)行了對(duì)比,從而可以更清晰的搞清楚每個(gè)概念的意義。
Elasticsearch |
MySQL |
index(索引,名詞) |
database |
doc type(文檔類(lèi)型) |
table |
document(文檔) |
row |
field(字段) |
column |
mapping(映射) |
schema |
query DSL(查詢(xún)語(yǔ)言) |
SQL |
然后,來(lái)看看倒排索引的概念(官方解釋?zhuān)5古潘饕撬阉饕娴幕彩荅lasticsearch能實(shí)現(xiàn)快速全文搜索的根本。歸納起來(lái),主要是對(duì)一個(gè)文檔內(nèi)容做兩步操作:分詞、建立“單詞-文檔”列表。舉個(gè)例子,假如有下面兩個(gè)文檔:
1 2 | 1. { "content" : "The quick brown fox jumped over the lazy dog" }
2. { "content" : "Quick brown foxes leap over lazy dogs in summer" }
|
Elasticsearch會(huì)使用分詞器對(duì)content字段的內(nèi)容進(jìn)行分詞,再根據(jù)單詞在文檔中是否出現(xiàn)建立如下所示的列表,√表示單詞在文檔中有出現(xiàn)。假如我們想搜索“quick brown”,只需要找到每個(gè)詞在哪個(gè)文檔中出現(xiàn)即可。如果有多個(gè)文檔匹配,可以根據(jù)匹配的程度進(jìn)行打分,找出相關(guān)性高的文檔。
Term |
Doc_1 |
Doc_2 |
Quick |
|
√ |
The |
√ |
|
brown |
√ |
√ |
dog |
√ |
|
dogs |
|
√ |
fox |
√ |
|
foxes |
|
√ |
in |
|
√ |
jumped |
√ |
|
lazy |
√ |
√ |
leap |
|
√ |
over |
√ |
√ |
quick |
√ |
|
summer |
|
√ |
the |
√ |
|
最后,我們?cè)倩剡^(guò)頭看看上面的映射的概念。類(lèi)似于MySQL在db schema中申明每個(gè)列的數(shù)據(jù)類(lèi)型、索引類(lèi)型等,Elasticsearch中使用mapping來(lái)做這件事。常用的是,在mapping中申明字段的數(shù)據(jù)類(lèi)型、是否建立倒排索引、建立倒排索引時(shí)使用什么分詞器。默認(rèn)情況下,Elasticsearch會(huì)為所有的string類(lèi)型數(shù)據(jù)使用standard分詞器建立倒排索引。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 | NOTE: 這里的 index 是blog,doc type是test
{
"blog" : {
"mappings" : {
"test" : {
"properties" : {
"activity_type" : {
"type" : "string" ,
"index" : "not_analyzed"
},
"address" : {
"type" : "string" ,
"analyzer" : "ik_smart"
},
"happy_party_id" : {
"type" : "integer"
},
"last_update_time" : {
"type" : "date" ,
"format" : "yyyy-MM-dd HH:mm:ss"
}
}
}
}
}
}</ index >
|
數(shù)據(jù)插入
在MySQL中,我們需要先建立database和table,申明db schema后才可以插入數(shù)據(jù)。而在Elasticsearch,可以直接插入數(shù)據(jù),系統(tǒng)會(huì)自動(dòng)建立缺失的index和doc type,并對(duì)字段建立mapping。因?yàn)榘虢Y(jié)構(gòu)化數(shù)據(jù)的數(shù)據(jù)結(jié)構(gòu)通常是動(dòng)態(tài)變化的,我們無(wú)法預(yù)知某個(gè)文檔中究竟有哪些字段,如果每次插入數(shù)據(jù)都需要提前建立index、type、mapping,那就失去了其作為NoSQL的優(yōu)勢(shì)了。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | {
"count" : 5,
"desc" : "hello world"
}
{
"blog" : {
"mappings" : {
"test" : {
"properties" : {
"count" : {
"type" : "long"
},
"desc" : {
"type" : "string"
}
}
}
}
}
}
|
然而這種靈活性是有限,比如上文我們提到,默認(rèn)情況下,Elasticsearch會(huì)為所有的string類(lèi)型數(shù)據(jù)使用standard分詞器建立倒排索引,那么如果某些字段不想建立倒排索引怎么辦。Elasticsearch提供了dynamic template的概念來(lái)針對(duì)一組index設(shè)置默認(rèn)mapping,只要index的名稱(chēng)匹配了,就會(huì)使用該template設(shè)置的mapping進(jìn)行字段映射。
下面所示即創(chuàng)建一個(gè)名稱(chēng)為blog的template,該template會(huì)自動(dòng)匹配以”blog_”開(kāi)頭的index,為其自動(dòng)建立mapping,對(duì)文檔中的所有string自動(dòng)增加一個(gè).raw字段,并且該字段不做索引。 這也是ELK中的做法,可以查看ELK系統(tǒng)中Elasticsearch的template,會(huì)發(fā)現(xiàn)有一個(gè)名為logstash的template。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 | {
"template" : "blog_*" ,
"mappings" : {
"_default_" : {
"dynamic_templates" : [{
"string_fields" : {
"mapping" : {
"type" : "string" ,
"fields" : {
"raw" : {
"index" : "not_analyzed" ,
"ignore_above" : 256,
"type" : "string"
}
}
},
"match_mapping_type" : "string"
}
}],
"properties" : {
"timestamp" : {
"doc_values" : true ,
"type" : "date"
}
},
"_all" : {
"enabled" : false
}
}
}
}
{
"count" : 5,
"desc" : "hello world"
}
|
插入問(wèn)題還有個(gè)話題就是批量插入。Elasticsearch提供了bulk API用來(lái)做批量的操作,你可以在該API中自由組合你要做的操作和數(shù)據(jù),一次性發(fā)送給Elasticsearch進(jìn)行處理,其格式是這樣的。
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 | action_and_meta_data\n
optional_source\n
action_and_meta_data\n
optional_source\n
....
action_and_meta_data\n
optional_source\n
比如:
{ "index" : { "_index" : "test" , "_type" : "type1" , "_id" : "1" } }
{ "field1" : "value1" }
{ "delete" : { "_index" : "test" , "_type" : "type1" , "_id" : "2" } }
{ "create" : { "_index" : "test" , "_type" : "type1" , "_id" : "3" } }
{ "field1" : "value3" }
{ "update" : { "_id" : "1" , "_type" : "type1" , "_index" : "test" } }
{ "doc" : { "field2" : "value2" } }
|
如果是針對(duì)相同的index和doc type進(jìn)行操作,則在REST API中指定index和type即可。批量插入的操作舉例如下:
1 2 3 4 5 6 7 8 9 10 11 | { "index" : {}}
{ "count" : 5, "desc" : "hello world 111" }
{ "index" : {}}
{ "count" : 6, "desc" : "hello world 222" }
{ "index" : {}}
{ "count" : 7, "desc" : "hello world 333" }
{ "index" : {}}
{ "count" : 8, "desc" : "hello world 444" }
|
數(shù)據(jù)查詢(xún)
Elasticsearch的查詢(xún)語(yǔ)法(query DSL)分為兩部分:query和filter,區(qū)別在于查詢(xún)的結(jié)果是要完全匹配還是相關(guān)性匹配。filter查詢(xún)考慮的是“文檔中的字段值是否等于給定值”,答案在“是”與“否”中;而query查詢(xún)考慮的是“文檔中的字段值與給定值的匹配程度如何”,會(huì)計(jì)算出每份文檔與給定值的相關(guān)性分?jǐn)?shù),用這個(gè)分?jǐn)?shù)對(duì)匹配了的文檔進(jìn)行相關(guān)性排序。
在實(shí)際使用中,要注意兩點(diǎn):第一,filter查詢(xún)要在沒(méi)有做倒排索引的字段上做,即上面mapping中增加的.raw字段;第二,通常使用filter來(lái)縮小查詢(xún)范圍,使用query進(jìn)行搜索,即二者配合使用。舉例來(lái)看,注意看三個(gè)不同查詢(xún)?cè)趯?xiě)法上的區(qū)別:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 | 1. 只使用query進(jìn)行查詢(xún):
查詢(xún)的結(jié)果是page_name字段中包含了wechat所有文檔
這里使用 size 來(lái)指定返回文檔的數(shù)量,默認(rèn)Elasticsearch是返回前100條數(shù)據(jù)的
{
"query" : {
"bool" : {
"must" : [{
"match" : {
"page_name" : "wechat"
}
},
{
"range" : {
"timestamp" : {
"gte" : 1481218631,
"lte" : 1481258231,
"format" : "epoch_second"
}
}
}]
}
},
"size" : 2
}
2. 只使用filter進(jìn)行查詢(xún):
查詢(xún)的結(jié)果是page_name字段值等于 "example.cn/wechat/view.html" 的所有文檔
{
"filter" : {
"bool" : {
"must" : [{
"term" : {
"page_name.raw" : "example.cn/wechat/view.html"
}
},
{
"range" : {
"timestamp" : {
"gte" : 1481218631,
"lte" : 1481258231,
"format" : "epoch_second"
}
}
}]
}
},
"size" : 2
}
3. 同時(shí)使用query與filter進(jìn)行查詢(xún):
查詢(xún)的結(jié)果是page_name字段值等于 "example.cn/wechat/view.html" 的所有文檔
{
"query" : {
"bool" : {
"filter" : [{
"bool" : {
"must" : [{
"term" : {
"page_name.raw" : "job.gikoo.cn/wechat/view.html"
}
},
{
"range" : {
"timestamp" : {
"gte" : 1481218631,
"lte" : 1481258231,
"format" : "epoch_second"
}
}
}]
}
}]
}
},
"size" : 2
}
|
聚合分析
類(lèi)似于MySQL中的聚合由分組和聚合計(jì)算兩種,Elasticsearch的聚合也有兩部分組成:Buckets與Metrics。Buckets相當(dāng)于SQL中的分組group by,而Metrics則相當(dāng)于SQL中的聚合函數(shù)COUNT,SUM,MAX,MIN等等。聚合分析自然離不開(kāi)對(duì)多個(gè)字段值進(jìn)行分組,在MySQL中,我們只要使用“group by c1, c2, c3”就可以完成這樣的功能,但是Elasticsearch沒(méi)有這樣的語(yǔ)法。Elasticsearch提供了另一種方法,即Buckets嵌套,仔細(xì)想想,似乎這種設(shè)計(jì)更加符合人的思維方式。舉例來(lái)看具體操作方法:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 | 1. 最簡(jiǎn)單的聚合查詢(xún)
為了簡(jiǎn)單,這里刪除了query的條件描述
將符合條件的文檔按照公司進(jìn)行聚合
這里有兩個(gè) size ,和aggs并列的 size =0表示返回結(jié)果不包含查詢(xún)結(jié)果,只返回聚合結(jié)果,terms里面的 size 表示返回的聚合結(jié)果數(shù)量
{
"aggs" : {
"company_terms" : {
"terms" : {
"field" : "company" ,
"size" : 2
}
}
},
"size" : 0
}
2. Buckets與Metric配合
將符合條件的文檔按照公司進(jìn)行聚合,并獲取每個(gè)公司最近一次操作的時(shí)間
{
"aggs" : {
"company_terms" : {
"terms" : {
"field" : "company" ,
"size" : 2
},
"aggs" : {
"latest_record" : {
"max" : {
"field" : "timestamp"
}
}
}
}
},
"size" : 0
}
3. Buckets嵌套
將符合條件的文檔先按照公司進(jìn)行聚合,再對(duì)每個(gè)公司下的門(mén)店進(jìn)行聚合,并獲取每個(gè)門(mén)店最近一次操作的時(shí)間
{
"aggs" : {
"company_terms" : {
"terms" : {
"field" : "company" ,
"size" : 1
},
"aggs" : {
"store_terms" : {
"terms" : {
"field" : "store" ,
"size" : 2
},
"aggs" : {
"latest_record" : {
"max" : {
"field" : "timestamp"
}
}
}
}
}
}
},
"size" : 0
}
|
|