# \[Elasticsearch\] Rechercher dans des array

**URL:** <https://discuss.elastic.co/t/elasticsearch-rechercher-dans-des-array/77141>\
**Category:** Discussions en français\
**Created:** [March 2, 2017, 1:08pm UTC](https://discuss.elastic.co/t/elasticsearch-rechercher-dans-des-array/77141 "2017-03-02T13:08:48Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![Benjamin\_Carriou](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/benjamin_carriou/32/47347_2.png) [@Benjamin\_Carriou](https://discuss.elastic.co/u/Benjamin_Carriou)\
**Post date:** [March 2, 2017, 1:08pm UTC](https://discuss.elastic.co/t/elasticsearch-rechercher-dans-des-array/77141/1 "2017-03-02T13:08:48Z")

</div>

Bonjour !

Actuellement, j'indexe des documents de la sorte::

```
{
   "took": 6,
   "timed_out": false,
   "_shards": {
	  "total": 5,
	  "successful": 5,
	  "failed": 0
   },
   "hits": {
	  "total": 3,
	  "max_score": 1,
	  "hits": [
		 {
			"_index": "nmb2b-2017.03.02",
			"_type": "FlightListByAerodromeReply",
			"_id": "AVqOakryqXFr_vbJPqLG",
			"_score": 1,
			"_source": {
			   "@timestamp": "2017-03-02T09:46:53.902Z",
			   "port": 18523,
			   "@version": "1",
			   "host": "192.168.10.195",
			   "parsed": {
				  "xmlns:S": "http://schemas.xmlsoap.org/soap/envelope/",
				  "Body": {
					 "FlightListByAerodromeReply": {
						"xmlns:common": "eurocontrol/cfmu/b2b/CommonServices",
						"requestReceptionTime": "2017-03-02 09:58:55",
						"data": {
						   "effectiveTrafficWindow": {
							  "wef": "2017-03-02 08:00",
							  "unt": "2017-03-02 20:00"
						   },
						   "flights": [
							  {
								 "flight": {
									"flightId": {
									   "keys": {
										  "aircraftId": "BEL32C",
										  "aerodromeOfDestination": "EBBR",
										  "estimatedOffBlockTime": "2017-03-02 07:45",
										  "airFiled": "false",
										  "nonICAOAerodromeOfDestination": "false",
										  "nonICAOAerodromeOfDeparture": "false",
										  "aerodromeOfDeparture": "LFPG"
									   },
									   "id": "AT02676815"
									}
								 }
							  },
							  {
								 "flight": {
									"flightId": {
									   "keys": {
										  "aircraftId": "AFR1033",
										  "aerodromeOfDestination": "LFPG",
										  "estimatedOffBlockTime": "2017-03-02 04:30",
										  "airFiled": "false",
										  "nonICAOAerodromeOfDestination": "false",
										  "nonICAOAerodromeOfDeparture": "false",
										  "aerodromeOfDeparture": "LGAV"
									   },
									   "id": "AT02680179"
									}
								 }
							  }
							
						},
						"xmlns:flight": "eurocontrol/cfmu/b2b/FlightServices",
						"requestId": "B2B_CUR:2419467",
						"xmlns:flow": "eurocontrol/cfmu/b2b/FlowServices",
						"xmlns:airspace": "eurocontrol/cfmu/b2b/AirspaceServices",
						"sendTime": "2017-03-02 09:59:42",
						"status": "OK"
					 }
				  }
			   },
			   "type": "FlightListByAerodromeReply",
			   "tags": []
			}
		 }
	  ]
   }
}

```

Comme vous pouvez le constater, je dispose d'un objet JSON 'flights' de type 'Array' qui peut contenir N 'flight':

```
"flights": [
  {
	 "flight": {
		"flightId": {
		   "keys": {
			  "aircraftId": "BEL32C",
			  "aerodromeOfDestination": "EBBR",
			  "estimatedOffBlockTime": "2017-03-02 07:45",
			  "airFiled": "false",
			  "nonICAOAerodromeOfDestination": "false",
			  "nonICAOAerodromeOfDeparture": "false",
			  "aerodromeOfDeparture": "LFPG"
		   },
		   "id": "AT02676815"
		}
	 }
  },

```

Or, dans certains cas, ces 'flight' peuvent ne pas contenir de champs 'id':

```
"flights": [
  {
	 "flight": {
		"flightId": {
		   "keys": {
			  "aircraftId": "FRA34Z",
			  "aerodromeOfDestination": "EBBR",
			  "estimatedOffBlockTime": "2017-03-02 07:45",
			  "airFiled": "false",
			  "nonICAOAerodromeOfDestination": "false",
			  "nonICAOAerodromeOfDeparture": "false",
			  "aerodromeOfDeparture": "LFPG"
		   }
		}
	 }
  },

```

Je souhaiterai donc effectuer une requête ES qui me permettrait de retourner tous les 'flight' qui contiennent bien un champs 'id'.

Etant novice dans les requêtes ES, pensez-vous que cela est possible ? Si oui comment faire ?

Merci d'avance ! 🙂

J'en profite pour mettre le mapping:

```
{
  "template": "nmb2b-*",
  "settings": { "number_of_shards": 5 },
  "mappings": {
		"FlightListByAerodromeReply": {
			"properties": {
			   "@timestamp": {
				  "type": "date",
				  "format": "strict_date_optional_time||epoch_millis"
			   },
			   "@version": {
				  "type": "string"
			   },
			   "host": {
				  "type": "string"
			   },
			   "parsed": {
				  "properties": {
					 "Body": {
						"properties": {
						   "FlightListByAerodromeReply": {
							  "properties": {
								 "data": {
									"properties": {
									   "effectiveTrafficWindow": {
										  "properties": {
											 "unt": {
												"type": "string"
											 },
											 "wef": {
												"type": "string"
											 }
										  }
									   },
									   "flights": {
										  "properties": {
											 "flight": {
												"properties": {
												   "flightId": {
													  "properties": {
														 "id": {
															"type": "string"
														 },
														 "keys": {
															"properties": {
															   "aerodromeOfDeparture": {
																  "type": "string"
															   },
															   "aerodromeOfDestination": {
																  "type": "string"
															   },
															   "airFiled": {
																  "type": "string"
															   },
															   "aircraftId": {
																  "type": "string"
															   },
															   "estimatedOffBlockTime": {
																  "type": "string"
															   },
															   "nonICAOAerodromeOfDeparture": {
																  "type": "string"
															   },
															   "nonICAOAerodromeOfDestination": {
																  "type": "string"
															   }
															}
														 }
													  }
												   }
												}
											 }
										  }
									   }
									}
								 },
								 "requestId": {
									"type": "string"
								 },
								 "requestReceptionTime": {
									"type": "string"
								 },
								 "sendTime": {
									"type": "string"
								 },
								 "status": {
									"type": "string"
								 },
								 "xmlns:airspace": {
									"type": "string"
								 },
								 "xmlns:common": {
									"type": "string"
								 },
								 "xmlns:flight": {
									"type": "string"
								 },
								 "xmlns:flow": {
									"type": "string"
								 }
							  }
						   }
						}
					 },
					 "xmlns:S": {
						"type": "string"
					 }
				  }
			   },
			   "port": {
				  "type": "long"
			   },
			   "query": {
				  "properties": {
					 "query_string": {
						"properties": {
						   "query": {
							  "type": "string"
						   }
						}
					 }
				  }
			   },
			   "type": {
				  "type": "string"
			   }
			}
		 }
  }
}
```

---

<div class="post-metadata">

**Author:** ![dadoonet](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/dadoonet/32/137187_2.png) [@dadoonet](https://discuss.elastic.co/u/dadoonet)\
**Post date:** [March 2, 2017, 6:59pm UTC](https://discuss.elastic.co/t/elasticsearch-rechercher-dans-des-array/77141/2 "2017-03-02T18:59:57Z")

</div>

Je pense que tu devrais indexer les flights indépendamment.  
Ca peut se faire en typant de type `nested` ton champ `flights`. Ainsi chaque flight sera indexé indépendamment des autres dans Lucene.

Ce qui devrait te permettre de faire une [nested query](https://www.elastic.co/guide/en/elasticsearch/reference/5.2/query-dsl-nested-query.html) avec une [exists query](https://www.elastic.co/guide/en/elasticsearch/reference/5.2/query-dsl-exists-query.html) sur ce champ `id`. Puis en demandant les [inner hits](https://www.elastic.co/guide/en/elasticsearch/reference/5.2/search-request-inner-hits.html), tu devrais pouvoir t'en sortir je pense.

Mais peut-être devrais-tu réfléchir à ton modèle et si ton besoin est de chercher des flights un par un et non des ensembles de flights regroupés, peut-être devrais-tu dénormaliser et indexer chaque flight individuellement?

---

<div class="post-metadata">

**Author:** ![Benjamin\_Carriou](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/benjamin_carriou/32/47347_2.png) [@Benjamin\_Carriou](https://discuss.elastic.co/u/Benjamin_Carriou)\
**Post date:** [March 3, 2017, 9:21am UTC](https://discuss.elastic.co/t/elasticsearch-rechercher-dans-des-array/77141/3 "2017-03-03T09:21:53Z")

</div>

Merci pour ta réponse.

Alors, j'ai modifié mon mapping en y ajoutant un type `nested` pour le champ `flights`:

```
{
  "order": 0,
  "template": "nmb2b-*",
  "settings": {
	"index": {
	  "number_of_shards": "5"
	}
  },
  "mappings": {
	"FlightListByAerodromeReply": {
	  "properties": {
		"@timestamp": {
		  "format": "strict_date_optional_time||epoch_millis",
		  "type": "date"
		},
		"port": {
		  "type": "long"
		},
		"query": {
		  "properties": {
			"query_string": {
			  "properties": {
				"query": {
				  "type": "string"
				}
			  }
			}
		  }
		},
		"@version": {
		  "type": "string"
		},
		"host": {
		  "type": "string"
		},
		"parsed": {
		  "properties": {
			"xmlns:S": {
			  "type": "string"
			},
			"Body": {
			  "properties": {
				"FlightListByAerodromeReply": {
				  "properties": {
					"xmlns:common": {
					  "type": "string"
					},
					"requestReceptionTime": {
					  "type": "string"
					},
					"data": {
					  "properties": {
						"effectiveTrafficWindow": {
						  "properties": {
							"wef": {
							  "type": "string"
							},
							"unt": {
							  "type": "string"
							}
						  }
						},
						"flights": {
						  "type": "nested", <== Ajout type 'nested'
						  "properties": {
							"flight": {
							  "properties": {
								"flightId": {
								  "properties": {
									"keys": {
									  "properties": {
										"aircraftId": {
										  "type": "string"
										},
										"aerodromeOfDestination": {
										  "type": "string"
										},
										"estimatedOffBlockTime": {
										  "type": "string"
										},
										"airFiled": {
										  "type": "string"
										},
										"nonICAOAerodromeOfDestination": {
										  "type": "string"
										},
										"nonICAOAerodromeOfDeparture": {
										  "type": "string"
										},
										"aerodromeOfDeparture": {
										  "type": "string"
										}
									  }
									},
									"id": {
									  "type": "string"
									}
								  }
								}
							  }
							}
						  }
						}
					  }
					},
					"xmlns:flight": {
					  "type": "string"
					},
					"requestId": {
					  "type": "string"
					},
					"xmlns:flow": {
					  "type": "string"
					},
					"xmlns:airspace": {
					  "type": "string"
					},
					"sendTime": {
					  "type": "string"
					},
					"status": {
					  "type": "string"
					}
				  }
				}
			  }
			}
		  }
		},
		"type": {
		  "type": "string"
		}
	  }
	}
  },
  "aliases": {}
}

```

J'ai ensuite essayé de faire une recherche, dans un premier temps, sans tester si le champ `id` existe afin de voir si les modifications fonctionnent. Voici la requête que j'essaye d'effectuer qui me permettrait de retrouver un `flight` en fonction de son `id` :

```
POST /nmb2b-2017.03.03/FlightListByAerodromeReply/_search
{
	"query": {
		"nested": {
			"path": "flights",
			"query": {
				"match": {
					"flights.flight.flightId.id": "AT02698286"
				}
			},
			"inner_hits": {}
		}
	}
}

```

Mais cette dernière me retourne l'erreur suivante:

```
{
   "error": {
	  "root_cause": [
		 {
			"type": "query_parsing_exception",
			"reason": "[nested] failed to find nested object under path [flights]",
			"index": "nmb2b-2017.03.03",
			"line": 4,
			"col": 21
		 }
	  ],
	  "type": "search_phase_execution_exception",
	  "reason": "all shards failed",
	  "phase": "query",
	  "grouped": true,
	  "failed_shards": [
		 {
			"shard": 0,
			"index": "nmb2b-2017.03.03",
			"node": "m7ZiycvQSeWaEslyS4QRWA",
			"reason": {
			   "type": "query_parsing_exception",
			   "reason": "[nested] failed to find nested object under path [flights]",
			   "index": "nmb2b-2017.03.03",
			   "line": 4,
			   "col": 21
			}
		 }
	  ]
   },
   "status": 400
}

```

Faut-il aussi typer les objets enfants de `flights` avec du `nested` ?

---

<div class="post-metadata">

**Author:** ![dadoonet](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/dadoonet/32/137187_2.png) [@dadoonet](https://discuss.elastic.co/u/dadoonet)\
**Post date:** [March 3, 2017, 10:47am UTC](https://discuss.elastic.co/t/elasticsearch-rechercher-dans-des-array/77141/4 "2017-03-03T10:47:32Z")

</div>

Tu as modifié un template apparement. As-tu supprimé l'index et recréé?

Quel est le mapping de l'index?

Idéalement, essaye d'envoyer des exemples beaucoup plus simples de façon à ce que cela soit plus facile à lire.

---

<div class="post-metadata">

**Author:** ![Benjamin\_Carriou](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/benjamin_carriou/32/47347_2.png) [@Benjamin\_Carriou](https://discuss.elastic.co/u/Benjamin_Carriou)\
**Post date:** [March 6, 2017, 8:23am UTC](https://discuss.elastic.co/t/elasticsearch-rechercher-dans-des-array/77141/5 "2017-03-06T08:23:56Z")

</div>

Après avoir ajouté un type `nested` aux champs enfants de `flights`, j'arrive à requeter:

```
POST /nmb2b-2017.03.03/FlightListByAerodromeReply/_search
{
	"query": {
		"nested": {
			"path": "parsed.Body.FlightListByAerodromeReply.data.flights",
			"query": {
				"nested": {
				   "path": "parsed.Body.FlightListByAerodromeReply.data.flights.flight",
				   "query": {
					   "nested": {
						  "path": "parsed.Body.FlightListByAerodromeReply.data.flights.flight.flightId",
						  "query": {
							  "match": {
								 "parsed.Body.FlightListByAerodromeReply.data.flights.flight.flightId.id": "AT02698286"
							}
						  }
					   }
				   }
				}
			},
			"inner_hits": {}
		}
	}
}

```

Mais le résultat que j'obtiens ne correspond pas à mes attentes.  
Comme tu l'as indiqué plus tôt, il serait plus judicieux de créer un document par `flight`.  
Si jamais cela peut intéresser quelqu'un, voilà comment j'ai procédé.

A l'aide de Logstash, je parse le document XML et je le `split` en plusieurs documents:

```
input {
		tcp {
				port => "7001"
				type => "FlightListByAerodromeReply"
		}
}
filter {
		if [type] == "FlightListByAerodromeReply" {
				xml {
						source => "message"
						#store_xml => false
						target => "parsed"
						#force_array => false
				}
				split {
						field => "parsed[Body]"
				}
				split {
						field => "parsed[Body][FlightListByAerodromeReply]"
				}
				split {
						field => "parsed[Body][FlightListByAerodromeReply][data]"
				}
				split {
						field => "parsed[Body][FlightListByAerodromeReply][data][effectiveTrafficWindow]"
				}
				split {
						field => "parsed[Body][FlightListByAerodromeReply][data][flights]"
				}
				split {
						field => "parsed[Body][FlightListByAerodromeReply][data][flights][flight]"
				}
				split {
						field => "parsed[Body][FlightListByAerodromeReply][data][flights][flight][flightId]"
				}
				split {
						field => "parsed[Body][FlightListByAerodromeReply][data][flights][flight][flightId][keys]"
				}
				mutate {
						add_field => { flightId => "%{parsed[Body][FlightListByAerodromeReply][data][flights][flight][flightId][id]}" }
						add_field => { aircraftId => "%{parsed[Body][FlightListByAerodromeReply][data][flights][flight][flightId][keys][aircraftId]}" }
						add_field => { aerodromeOfDestination => "%{parsed[Body][FlightListByAerodromeReply][data][flights][flight][flightId][keys][aerodromeOfDestination]}" }
						add_field => { estimatedOffBlockTime => "%{parsed[Body][FlightListByAerodromeReply][data][flights][flight][flightId][keys][estimatedOffBlockTime]}" }
						add_field => { airFiled => "%{parsed[Body][FlightListByAerodromeReply][data][flights][flight][flightId][keys][airFiled]}" }
						add_field => { nonICAOAerodromeOfDestination => "%{parsed[Body][FlightListByAerodromeReply][data][flights][flight][flightId][keys][nonICAOAerodromeOfDestination]}" }
						add_field => { nonICAOAerodromeOfDeparture => "%{parsed[Body][FlightListByAerodromeReply][data][flights][flight][flightId][keys][nonICAOAerodromeOfDeparture]}" }
						add_field => { aerodromeOfDeparture => "%{parsed[Body][FlightListByAerodromeReply][data][flights][flight][flightId][keys][aerodromeOfDeparture]}" }
						add_field => { wef => "%{parsed[Body][FlightListByAerodromeReply][data][effectiveTrafficWindow][wef]}" }
						add_field => { unt => "%{parsed[Body][FlightListByAerodromeReply][data][effectiveTrafficWindow][unt]}" }
						add_field => { requestReceptionTime => "%{parsed[Body][FlightListByAerodromeReply][requestReceptionTime]}" }
						add_field => { sendTime => "%{parsed[Body][FlightListByAerodromeReply][sendTime]}" }
						add_field => { status => "%{parsed[Body][FlightListByAerodromeReply][status]}" }
						remove_field => ["message"]
						remove_field => ["parsed"]
				}
		}
}

```

Ce qui me permet d'obtenir, une fois le tout indexé:

```
POST /nmb2b-2017.03.03/FlightListByAerodromeReply/_search

	{
   "took": 3,
   "timed_out": false,
   "_shards": {
	  "total": 5,
	  "successful": 5,
	  "failed": 0
   },
   "hits": {
	  "total": 18863,
	  "max_score": 1,
	  "hits": [
		 {
			"_index": "nmb2b-2017.03.03",
			"_type": "FlightListByAerodromeReply",
			"_id": "AVqU5s__qXFr_vbJPt1p",
			"_score": 1,
			"_source": {
			   "requestReceptionTime": "2017-03-03 16:12:34",
			   "flightId": "AT02698316",
			   "airFiled": "false",
			   "type": "FlightListByAerodromeReply",
			   "tags": [],
			   "sendTime": "2017-03-03 16:12:35",
			   "aircraftId": "TAY123G",
			   "@timestamp": "2017-03-03T16:00:32.479Z",
			   "port": 36157,
			   "aerodromeOfDestination": "LFBO",
			   "estimatedOffBlockTime": "2017-03-03 00:15",
			   "@version": "1",
			   "host": "192.168.10.195",
			   "wef": "2017-03-03 00:00",
			   "unt": "2017-03-03 23:59",
			   "nonICAOAerodromeOfDestination": "false",
			   "nonICAOAerodromeOfDeparture": "false",
			   "aerodromeOfDeparture": "LFPG",
			   "status": "OK"
			}
		 },
		 {
			"_index": "nmb2b-2017.03.03",
			"_type": "FlightListByAerodromeReply",
			"_id": "AVqU5s__qXFr_vbJPt1q",
			"_score": 1,
			"_source": {
			   "requestReceptionTime": "2017-03-03 16:12:34",
			   "flightId": "AT02697997",
			   "airFiled": "false",
			   "type": "FlightListByAerodromeReply",
			   "tags": [],
			   "sendTime": "2017-03-03 16:12:35",
			   "aircraftId": "FPO701",
			   "@timestamp": "2017-03-03T16:00:32.479Z",
			   "port": 36157,
			   "aerodromeOfDestination": "LFML",
			   "estimatedOffBlockTime": "2017-03-03 00:35",
			   "@version": "1",
			   "host": "192.168.10.195",
			   "wef": "2017-03-03 00:00",
			   "unt": "2017-03-03 23:59",
			   "nonICAOAerodromeOfDestination": "false",
			   "nonICAOAerodromeOfDeparture": "false",
			   "aerodromeOfDeparture": "LFPG",
			   "status": "OK"
			}
		 }

```

Merci pour ton aide et à bientôt !  
Cordialement, Benjamin

---

<div class="post-metadata">

**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\
**Post date:** [April 3, 2017, 8:24am UTC](https://discuss.elastic.co/t/elasticsearch-rechercher-dans-des-array/77141/6 "2017-04-03T08:24:38Z")

</div>

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.