Retrieving a record To get the record out of Elasticsearch, we use a similar _index, _type, and _id, however the HTTP verb adjustments to GET: GET /website/blog/123? lovely The reaction comprises the by-now-familiar metadata parts, plus the _source box, which includes the unique JSON record that we despatched to Elasticsearch once we listed it: { "_index" : "website", "_type" : "blog", "_id" : "123", "_version" : 1, "found" : real, "_source" : { "title": "My first weblog entry", "text": "Just attempting this out... ", "date": "2014/01/01" } } be aware including beautiful to the query-string parameters for any request, as within the previous instance, factors Elasticsearch to pretty-print the JSON reaction to make it extra readable. The _source box, although, isn’t pretty-printed. in its place we come again the exact same JSON string that we handed in. The reaction to the GET request comprises {"found": true}. This confirms that the record used to be came upon. If we have been to request a rfile that doesn’t exist, we'd nonetheless get a JSON reaction, yet came upon will be set to fake. additionally, the HTTP reaction code will be 404 now not came across rather than two hundred okay. we will see this by means of passing the -i argument to curve, which motives it to reveal the reaction headers: curl -i -XGET http://localhost:9200/website/blog/124? lovely The reaction now feels like this: HTTP/1. 1 404 now not discovered Content-Type: application/json; charset=UTF-8 Content-Length: eighty three { "_index" : "website", "_type" : "blog", "_id" : "124", "found" : fake } Retrieving a part of a rfile through default, a GET request will go back the full rfile, as saved within the _source box. yet might be all you have an interest in is the identify box. person fields may be asked by utilizing the _source parameter. a number of fields might be laid out in a comma-separated record: GET /website/blog/123? _source=title,text The _source box now comprises simply the fields that we asked and has filtered out the date box: { "_index" : "website", "_type" : "blog", "_id" : "123", "_version" : 1, "exists" : actual, "_source" : { "title": "My first weblog access" , "text": "Just making an attempt this out... " } } Or if you'd like simply the _source box with none metadata, you should use the _source endpoint: GET /website/blog/123/_source which returns simply the next: { "title": "My first weblog entry", "text": "Just making an attempt this out... ", "date": "2014/01/01" } Checking even if a rfile Exists If all you must do is to examine even if a record exists—you’re now not drawn to the content material at all—then use the top process rather than the GET strategy. HEAD requests don’t go back a physique, simply HTTP headers: curl -i -XHEAD http://localhost:9200/website/blog/123 Elasticsearch will go back a 2 hundred okay prestige code if the rfile exists: HTTP/1. 1 2 hundred okay Content-Type: text/plain; charset=UTF-8 Content-Length: zero And a 404 no longer came upon if it doesn’t exist: curl -i -XHEAD http://localhost:9200/website/blog/124 HTTP/1. 1 404 no longer came across Content-Type: text/plain; charset=UTF-8 Content-Length: zero in fact, simply because a rfile didn’t exist in the event you checked it, doesn’t suggest that it won’t exist a millisecond later: one other method may well create the rfile meanwhile.