? 若在傳統DBMS 關系型資料庫中查詢海量資料，特別是模糊查詢，一般我們都是使用like %查詢的值%，但這樣會導致無法應用索引，從而形成全表掃描效率低下，即使是在有索引的欄位精確值查找，面對海量資料，效率也是相對較低的，所以目前一般的互聯網公司或大型公司，若要查詢海量資料，最好的辦法就是使用搜索引擎，目前比較主流的搜索引擎框架就是：Elasticsearch，故今天我這里總結了Elasticsearch必知必會的干貨知識一：ES索引檔案的CRUD，后面陸續還會有其它干貨知識分享，敬請期待，
-
ES索引檔案的CRUD（6.X與7.X有區別，6.X中支持一個index創建多個type，而7.X中及以上只支持1個固定的type，即：_doc，API用法上也稍有不同）：
-
Create創建索引檔案【POST index/type/id可選，如果index、type、id已存在則重建索引檔案（先洗掉后創建索引檔案，與Put index/type/id 原理相同），如果在指定id情況下需要限制自動更新，則可以使用：index/type/id?op_type=create 或 index/type/id/_create，指明操作型別為創建，這樣當存在的記錄的情況下會報錯】
POST demo_users/_doc 或 demo_users/_doc/2vJKsm8BriJODA6s9GbQ/_createRequest Body:
{ "userId":1, "username":"張三", "role":"administrator", "enabled":true, "createdDate":"2020-01-01T12:00:00" }Response Body:
{ "_index": "demo_users", "_type": "_doc", "_id": "2vJKsm8BriJODA6s9GbQ", "_version": 1, "result": "created", "_shards": { "total": 2, "successful": 1, "failed": 0 }, "_seq_no": 0, "_primary_term": 1 } -
Get獲取索引檔案【Get index/type/id】
Get demo_users/_doc/123Response Body:
{ "_index": "demo_users", "_type": "_doc", "_id": "123", "_version": 1, "found": true, "_source": { "userId": 1, "username": "張三", "role": "administrator", "enabled": true, "createdDate": "2020-01-01T12:00:00" } } -
Index Put重建索引檔案【PUT index/type/id 或 index/type/id?op_type=index，id必傳，如果id不存在檔案則創建檔案，否則先洗掉原有id檔案后再重新創建檔案，version加1】
Put/POST demo_users/_doc/123 或 demo_users/_doc/123?op_type=indexRequest Body:
{ "userId":1, "username":"張三", "role":"administrator", "enabled":true, "createdDate":"2020-01-01T12:00:00", "remark":"僅演示" }Response Body:
{ "_index": "demo_users", "_type": "_doc", "_id": "123", "_version": 4, "result": "updated", "_shards": { "total": 2, "successful": 2, "failed": 0 }, "_seq_no": 10, "_primary_term": 1 } -
Update更新索引檔案【POST index/type/id/_update 請求體必需是{"doc":{具體的檔案JSON}},如果指定的鍵欄位已存在則更新，如果指定的鍵欄位不存在則附加新的鍵值對，支持多層級嵌套，多次請求，如果有欄位值有更新則version加1，否則提示更新0條 】
POST demo_users/_doc/123/_updateRequest Body:
{ "doc": { "userId": 1, "username": "張三", "role": "administrator", "enabled": true, "createdDate": "2020-01-01T12:00:00", "remark": "僅演示POST更新5", "updatedDate": "2020-01-17T15:30:00" } }Response Body:
{ "_index": "demo_users", "_type": "_doc", "_id": "123", "_version": 26, "result": "updated", "_shards": { "total": 2, "successful": 2, "failed": 0 }, "_seq_no": 35, "_primary_term": 1 } -
Delete洗掉索引檔案【DELETE index/type/id】
DELETE demo_users/_doc/123Response Body:
{ "_index": "demo_users", "_type": "_doc", "_id": "123", "_version": 2, "result": "deleted", "_shards": { "total": 2, "successful": 2, "failed": 0 }, "_seq_no": 39, "_primary_term": 1 } -
Bulk批量操作檔案【
POST _bulk 或 index/_bulk 或 index/type/_bulk一次請求支持進行多個索引、多個type的多種不同的CRUD操作，如果操作中有某個出現錯誤不會影響其它操作；】POST _bulkRequest Body:(注意最后還得多一個換行，因為ES是根據換行符來識別多條命令的，如果缺少最后一潭訓行則會報錯，注意請求體非標準的JSON，每行才是一個JSON，整體頂多可看成是\n區分的JSON物件陣列)
{ "index" : { "_index" : "demo_users_test", "_type" : "_doc", "_id" : "1" } } { "bulk_field1" : "測驗創建index" } { "delete" : { "_index" : "demo_users", "_type" : "_doc", "_id" : "123" } } { "create" : { "_index" : "demo_users", "_type" : "_doc", "_id" : "2" } } { "bulk_field2" : "測驗創建index2" } { "update" : { "_index" : "demo_users_test","_type" : "_doc","_id" : "1" } } { "doc": {"bulk_field1" : "測驗創建index1","bulk_field2" : "測驗創建index2"} }Response Body:
{ "took": 162, "errors": true, "items": [ { "index": { "_index": "demo_users_test", "_type": "_doc", "_id": "1", "_version": 8, "result": "updated", "_shards": { "total": 2, "successful": 2, "failed": 0 }, "_seq_no": 7, "_primary_term": 1, "status": 200 } }, { "delete": { "_index": "demo_users", "_type": "_doc", "_id": "123", "_version": 2, "result": "not_found", "_shards": { "total": 2, "successful": 2, "failed": 0 }, "_seq_no": 44, "_primary_term": 1, "status": 404 } }, { "create": { "_index": "demo_users", "_type": "_doc", "_id": "2", "status": 409, "error": { "type": "version_conflict_engine_exception", "reason": "[_doc][2]: version conflict, document already exists (current version [1])", "index_uuid": "u7WE286CQnGqhHeuwW7oyw", "shard": "2", "index": "demo_users" } } }, { "update": { "_index": "demo_users_test", "_type": "_doc", "_id": "1", "_version": 9, "result": "updated", "_shards": { "total": 2, "successful": 2, "failed": 0 }, "_seq_no": 8, "_primary_term": 1, "status": 200 } } ] } -
mGet【POST
_mget 或 index/_mget 或 index/type/_mget，如果指定了index或type，則請求報文中則無需再指明index或type，可以通過_source指明要查詢的include以及要排除exclude的欄位】POST _mgetRequest Body:
{ "docs": [ { "_index": "demo_users", "_type": "_doc", "_id": "12345" }, { "_index": "demo_users", "_type": "_doc", "_id": "1234567", "_source": [ "userId", "username", "role" ] }, { "_index": "demo_users", "_type": "_doc", "_id": "1234", "_source": { "include": [ "userId", "username" ], "exclude": [ "role" ] } } ] }Response Body:
{ "docs":[ { "_index":"demo_users", "_type":"_doc", "_id":"12345", "_version":1, "found":true, "_source":{ "userId":1, "username":"張三", "role":"administrator", "enabled":true, "createdDate":"2020-01-01T12:00:00" } }, { "_index":"demo_users", "_type":"_doc", "_id":"1234567", "_version":7, "found":true, "_source":{ "role":"administrator", "userId":1, "username":"張三" } }, { "_index":"demo_users", "_type":"_doc", "_id":"1234", "_version":1, "found":true, "_source":{ "userId":1, "username":"張三" } } ] }POST demo_users/_doc/_mgetRequest Body:
{ "ids": [ "1234", "12345", "123457" ] }Response Body:
{ "docs":[ { "_index":"demo_users", "_type":"_doc", "_id":"1234", "_version":1, "found":true, "_source":{ "userId":1, "username":"張三", "role":"administrator", "enabled":true, "createdDate":"2020-01-01T12:00:00", "remark":"僅演示" } }, { "_index":"demo_users", "_type":"_doc", "_id":"12345", "_version":1, "found":true, "_source":{ "userId":1, "username":"張三", "role":"administrator", "enabled":true, "createdDate":"2020-01-01T12:00:00" } }, { "_index":"demo_users", "_type":"_doc", "_id":"123457", "found":false } ] } -
_update_by_query根據查詢條件更新匹配到的索引檔案的指定欄位【
POST index/_update_by_query請求體寫查詢條件以及更新的欄位，更新欄位這里采用了painless腳本進行靈活更新】POST demo_users/_update_by_queryRequest Body:（意思是查詢role=administrator【可能大家看到keyword，這是因為role欄位為text型別，無法直接匹配，需要借助于子欄位role.keyword，如果有不理解后面會有簡要說明】，更新role為poweruser、remark為remark+采用_update_by_query更新）
{ "script":{ "source":"ctx._source.role=params.role;ctx._source.remark=ctx._source.remark+params.remark", "lang":"painless", "params":{ "role":"poweruser", "remark":"采用_update_by_query更新" } }, "query":{ "term":{ "role.keyword":"administrator" } } }painless寫法請具體參考：painless語法教程
Response Body:
{ "took": 114, "timed_out": false, "total": 6, "updated": 6, "deleted": 0, "batches": 1, "version_conflicts": 0, "noops": 0, "retries": { "bulk": 0, "search": 0 }, "throttled_millis": 0, "requests_per_second": -1, "throttled_until_millis": 0, "failures": [ ] } -
_delete_by_query根據查詢條件洗掉匹配到的索引檔案【
POST index/_delete_by_query請求體寫查詢匹配條件】POST demo_users/_delete_by_queryRequest Body:（意思是查詢enabled=false）
{ "query": { "match": { "enabled": false } } }Response Body:
{ "took":29, "timed_out":false, "total":3, "deleted":3, "batches":1, "version_conflicts":0, "noops":0, "retries":{ "bulk":0, "search":0 }, "throttled_millis":0, "requests_per_second":-1, "throttled_until_millis":0, "failures":[ ] } -
search查詢
-
URL GET查詢（GET index/_search?q=query_string語法，注意中文內容默認分詞器是一個漢字拆分成一個term）
A.Term Query:【即分詞片段(詞條)查詢，注意這里講的包含是指與分詞片段匹配】 GET /demo_users/_search?q=role:poweruser //指定欄位查詢,即：欄位包含查詢的值 GET /demo_users/_search?q=poweruser //泛查詢(沒有指定查詢的欄位)，即查詢檔案中所有欄位包含poweruser的值，只要有一個欄位符合，那么該檔案將會被回傳 B.Phrase Query【即分組查詢】 運算子有：AND / OR / NOT 或者表示為： && / || / ! +表示must -表示must_not 例如：field:(+a -b)意為field中必需包含a但不能包含b GET /demo_users/_search?q=remark:(POST test) GET /demo_users/_search?q=remark:(POST OR test) GET /demo_users/_search?q=remark:"POST test" //分組查詢，即：查詢remark中包含POST 或 test的檔案記錄 GET /demo_users/_search?q=remark:(test AND POST) //remark同時包含test與POST GET /demo_users/_search?q=remark:(test NOT POST) //remark包含test但不包含POST C.范圍查詢 區間表示：[]閉區間，{}開區間 如：year:[2019 TO 2020] 或 {2019 TO 2020} 或 {2019 TO 2020] 或 [* TO 2020] 算數符號 year:>2019 或 (>2012 && <=2020) 或 (+>=2012 +<=2020) GET /demo_users/_search?q=userId:>123 //查詢userId欄位大于123的檔案記錄 D.通配符查詢 ?表示匹配任意1個字符，*表示匹配0或多個字符 例如：role:power* , role:use? GET /demo_users/_search?q=role:power* //查詢role欄位前面是power，后面可以是0或多個其它任意字符， 可使用正則運算式，如：username:張三\d+ 可使用近似查詢偏移量（slop）提高查詢匹配結果【使用~N，N表示偏移量】 GET /demo_users/_search?q=remark:tett~1 //查詢remark中包含test的檔案，但實際寫成了tett，故使用~1偏移近似查詢，可以獲得test的查詢結果 GET /demo_users/_search?q=remark:"i like shenzhen"~2 //查詢i like shenzhen但實際remark欄位中值為：i like hubei and shenzhen，比查詢值多了 hubei and，這里使用~2指定可偏移相隔2個term（這里即兩個單詞），最終也是可以查詢出結果 -
DSL POST查詢（POST index/_search）
POST demo_users/_searchRequest Body:
{ "query":{ "bool":{ "must":[ { "term":{ "enabled":"true" #查詢enabled=true } }, { "term":{ "role.keyword":"poweruser" #且role=poweruser } }, { "query_string":{ "default_field":"username.keyword", "query":"張三" #且 username 包含張三 } } ], "must_not":[ ], "should":[ ] } }, "from":0, "size":1000, "sort":[ { "createdDate":"desc" #根據createdDate倒序 } ], "_source":{ #指明回傳的欄位，includes需回傳欄位，excludes不需要回傳欄位 "includes":[ "role", "username", "userId", "remark" ], "excludes":[ ] } }
-
-
具體用法可參見：
【Elasticsearch】query_string的各種用法
Elasticsearch中 match、match_phrase、query_string和term的區別
Elasticsearch Query DSL 整理總結
[布爾查詢Bool Query]
最后附上ES官方的API操作鏈接指引：
Indices APIs：負責索引Index的創建（create）、洗掉（delete）、獲取（get）、索引存在（exist）等操作，
Document APIs：負責索引檔案的創建（index）、洗掉（delete）、獲取（get）等操作，
Search APIs：負責索引檔案的search（查詢），Document APIS根據doc_id進行查詢，Search APIs]根據條件查詢，
Aggregations：負責針對索引的檔案各維度的聚合（Aggregation），
cat APIs：負責查詢索引相關的各類資訊查詢，
Cluster APIs：負責集群相關的各類資訊查詢，
轉載請註明出處，本文鏈接：https://www.uj5u.com/shujuku/35403.html
標籤：大數據
上一篇：hive內置方法一覽
下一篇：hadoop學習摘要