Basic usage of elasticsearch

Regarding the background of elasticsearch and the concept of distinguishing from conventional relational databases

One, install elasticsearch

brew install elasticsearch

If an error is reported, follow the prompts to install java related environment first. You can basically give up without a ladder.

Two, install kibana

1. brew install kibana
2. brew services start kibana
3. Browser access http://localhost:5601-> Dev Tools

Three, index index

1. Create an index named my_index PUT my_index
2. View index information GET my_index/_settings
3. Delete index DELETE my_index

Four, type

1. Only one can be created in the 6+ version, and no longer can be created in the 7+ version (a default one will be given).
2. In the introduction of this article, there is no need to create type specially

Five, document Document

1. Create data rows

   PUT my_index/my_type/1
   {
         
         
     "id":1,
     "name":"Alice",
     "date":"2020-01-01",
     "like": "apple and orange"
   }
   

   Where my_index is the index, my_type is the type, and 1 is the document id. The request method can be PUTor the POST
   document id can be specified, or it can be automatically generated, it must be usedPOST

   POST my_index/my_type
   {
         
         
     "id":2,
     "name":"Bob",
     "date":"2020-01-01",
     "like":"banana and pear"
   }
   

   Take a look GET my_index/_searchand observe the _id field. Bob's is AXBXHlWdGy6roZEZWR9P

2. View the data row.
   Known id, view Bob's information separatelyGET my_index/my_type/AXBXHlWdGy6roZEZWR9P

   {
         
         
    "_index": "my_index",
    "_type": "my_type",
    "_id": "AXBXHlWdGy6roZEZWR9P",
    "_version": 1,
    "found": true,
    "_source": {
         
         
      "id": 1102,
      "name": "Bob",
      "date": "2020-01-02"
    }
   }
   

   See only rows of data content GET my_index/my_type/AXBXHlWdGy6roZEZWR9P/_source
   to view only id and name fields contentGET my_index/my_type/AXBXHlWdGy6roZEZWR9P/_source?_source=id,name

3. The update of the data row
   elasticsearch is to mark the old document as deleted (deleted at a later point in time) and inaccessible, create a new document and index it.
   (1) Update the entire row of data in the same way as creating.
   (2) Update a field individually

   PUT my_index/my_type/1/_update
   {
         
         
     "doc":{
         
         
       "id":1103
     }
   }
   

Six, result analysis

GET my_index/_search

{
    
    
  "took": 0,
  "timed_out": false,
  "_shards": {
    
    
    "total": 5,
    "successful": 5,
    "skipped": 0,
    "failed": 0
  },
  "hits": {
    
    
    "total": 2,
    "max_score": 1,
    "hits": [
      {
    
    
        "_index": "my_index",
        "_type": "my_type",
        "_id": "AXBXHlWdGy6roZEZWR9P",
        "_score": 1,
        "_source": {
    
    
          "id": 1102,
          "name": "Bob",
          "date": "2020-01-02",
          "like": "banana and pear"
        }
      },
      {
    
    
        "_index": "my_index",
        "_type": "my_type",
        "_id": "1",
        "_score": 1,
        "_source": {
    
    
          "id": 1101,
          "name": "Alice",
          "date": "2020-01-01",
          "like": "apple and orange"
        }
      }
    ]
  }
}

1. took
   the number of milliseconds the entire search request took.
2. timed_out
   query timed out or not. You may be limited GET /_search?timeout=10msor GET /_search?timeout=1sthe return of the results collected before the request times out. (It should be noted that timeout will not stop the execution of the query, it only tells you the node that has successfully returned the result and then closes the connection. In the background, other shards may still execute the query, although the result has been sent)
3. _shards The
   number of shards participating in the query (total field), how many are successful (successful field), and how many are failed (failed field)
4. The hits
   total field represents the total number of matched documents;
   max_score refers to the maximum value of _score in all document matching queries; the
   hits array contains the first 10 matched data by default. Each node has a _score field, which is the relevance score, which measures how well the document matches the query. The array is sorted in descending order of _score. It can be used sizeand fromset the number and the offset return resultsGET my_index/_search?size=1&from=1

Seven, search-simple query string

It is not recommended to directly expose the query string search to users, unless these users are trustworthy with your data and cluster.

1. Field like contains apple GET my_index/my_type/_search?q=like:apple
2. The field name contains Alice or the field id is 1102 GET my_index/_search?q=+name:Alice +id:1102
3. The field name contains Alice or the field id is not 1103 GET my_index/_search?q=+name:Alice -id:1103
4. The field name contains Alice or contains Bob GET my_index/my_type/_search?q=name:(Alice Bob)
5. Field date >= '2020-01-02' GET my_index/my_type/_search?q=date:>=2020-01-02

Eight, search-structured query

GET is better for submitting query requests, but because GET carrying interactive data is not well supported, the search API also supports POST.

1. term filtering. Used to match exactly which values, such as numbers, dates, boolean values ​​or not_analyzed strings.

   POST my_index/my_type/_search 
   {
         
          
     "query": {
         
         
       "term": {
         
         "id": 1102}
     }
   }
   

2. terms filtering. Similar to term, the field specifies multiple values, which is equivalent to in.

   POST my_index/my_type/_search 
   {
         
          
     "query": {
         
         
       "terms": {
         
          "id": [1102, 1103]}
     }
   }
   

3. range filtering. Range lookup.

   POST my_index/my_type/_search 
   {
         
          
     "query": {
         
         
       "range": {
         
         
         "date": {
         
         
           "gte": "2020",
           "lt": "2020-01-02"
         }
       }
     }
   }
   

4. match query When
   doing an exact match search, it is best to use a filter statement, because the filter statement can cache data.

   POST my_index/my_type/_search 
   {
         
          
     "query": {
         
         
       "match": {
         
         "like": "apple"}
     }
   }
   

5. multi_match query

   POST my_index/my_type/_search 
   {
         
          
     "query": {
         
         
       "multi_match": {
         
         "like": ["apple", "banana"]}
     }
   }
   

6. Other
   size and from

   POST my_index/my_type/_search
   {
         
          
     "query": {
         
         },
     "size": 1,
     "from": 0
   }
   

   Return only specific fields

   POST my_index/my_type/_search
   {
         
          
     "query": {
         
         },
     "_source": ["id", "name"]
   }		
   

Three, install the ik Chinese word segmenter

1. Through the elasticsearch-plugininstallation, note that the elasticsearch version number corresponds to the ik version number, the version number of this installation is 5.6.15
   /usr/local/Cellar/[email protected]/5.6.15/bin/elasticsearch-plugin install https://github.com/medcl/elasticsearch-analysis-ik/releases/download/v5.6.15/elasticsearch-analysis-ik-5.6.15.zip

2. Stop elasticsearch and restart it using startup method 2. Find the line that loads ik in the startup information
   

3. brew info elasticsearchCheck the relevant information directory, find
   Config: /usr/local/etc/elasticsearch/ and
   enter the directory cd /usr/local/etc/elasticsearch/analysis-ik, where the IKAnalyzer.cfg.xml file is the configuration dictionary file