如何使用Elasticsearch提取所有文档中嵌套字段color1的全部值

Hey there! Let's figure out how to get all the color1 values mapped to each website in your Elasticsearch documents. The reason your match_all with size:0 didn't work is that size:0 just tells ES not to return raw document hits—you need to use aggregations to group and extract the data you want. Here are a couple of tailored solutions depending on your exact needs:

1. Group by Website and Extract color1 Values (Ideal for Aggregated Results) #

If you want to see each unique website paired with its associated color1 values, use a combination of terms aggregation (to group websites) and either top_hits or nested aggregation (to handle the color array):

Option A: For Single Color Entry Per Website

If each website's document only has one color object in the color array, this simple query works perfectly:

GET /your_index_name/_search
{
  "size": 0, // Skip returning raw docs, focus only on aggregations
  "aggs": {
    "grouped_websites": {
      "terms": {
        "field": "website.keyword", // Use .keyword to avoid text field tokenization issues
        "size": 10000 // Adjust this number to cover all your unique websites
      },
      "aggs": {
        "extract_color1": {
          "top_hits": {
            "size": 1,
            "_source": {
              "includes": ["color.color1"]
            }
          }
        }
      }
    }
  }
}

Option B: For Multiple Color Entries Per Website

If a website has multiple documents or multiple color objects in the array, use a nested aggregation to capture all unique color1 values for each website:

GET /your_index_name/_search
{
  "size": 0,
  "aggs": {
    "grouped_websites": {
      "terms": {
        "field": "website.keyword",
        "size": 10000
      },
      "aggs": {
        "color_nested": {
          "nested": {
            "path": "color" // Target the nested color array to iterate through each object
          },
          "aggs": {
            "all_color1_values": {
              "terms": {
                "field": "color.color1.keyword",
                "size": 10000
              }
            }
          }
        }
      }
    }
  }
}

2. Get All Website-color1 Pairs Directly (No Aggregation) #

If you just want a flat list of every document's website and corresponding color1, use _source filtering to only return the fields you care about, and set a large enough size to fetch all documents (use scroll/search_after if you have more than 10k docs):

GET /your_index_name/_search
{
  "size": 10000, // Adjust based on your total document count
  "_source": {
    "includes": ["website", "color.color1"]
  }
}

Just replace your_index_name with the actual name of your Elasticsearch index, and tweak the size values to make sure you capture all your data.

内容的提问来源于stack exchange，提问作者Victoire