Welcome to OGeek Q&A Community for programmer and developer-Open, Learning and Share
Welcome To Ask or Share your Answers For Others

Categories

0 votes
1.6k views
in Technique[技术] by (71.8m points)

json - How to 'join' two indexes in ElasticSearch

I have two indexes that must be separated:

// index = `order_item`
{
    "ID": 1,
    "Name": "Shoes",
    "Price": 9.99,
    "OrderID": 82
},{
    "ID": 1,
    "Name": "Hat",
    "Price": 19.99,
    "OrderID": 82
}

// index = `order`
{
    "ID": 82,
    "Customer": "John Smith"
}

How would I 'join' these two tables on a search, such that it would return something along the lines of:

results = {
    "ID": 1,
    "Name": "Shoes",
    "Price": 9.99,
    "Order.ID": 82,
    "Customer": "John Smith"
},{
    "ID": 1,
    "Name": "Hat",
    "Price": 19.99,
    "Order.ID": 82,
    "Customer": "John Smith"
}
See Question&Answers more detail:os

与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
Welcome To Ask or Share your Answers For Others

1 Reply

0 votes
by (71.8m points)

As answered in your other question, nothing prevents you from storing the Customer name inside each order_item document at indexing time, while still having a dedicated index orders also containing the Customer data. Remember that it's all about cleverly denormalizing your data so that each of your documents be as "self-contained" as you need.

curl -XPUT localhost:9200/order_items/order_item/1 -d '{
    "ID": 1,
    "Name": "Shoes",
    "Price": 9.99,
    "OrderID": 82,
    "Customer": "John Smith"
}'

curl -XPUT localhost:9200/order_items/order_item/2 -d '{
    "ID": 2,
    "Name": "Hat",
    "Price": 19.99,
    "OrderID": 82,
    "Customer": "John Smith"
}

The advantages of this solution is that each order item is completely self-contained, and you can group/aggregate them on OrderID in order to get all items of a given order.

Also, as @JohnAment mentioned in his comment, the order/order_item use case is also a good candidate for using either

  1. the parent/child relationship
  2. or nested objects.

In the first case, you'd have one order "parent" document...

curl -XPUT localhost:9200/orders/order/82 -d '{
    "ID": 82,
    "Customer": "John Smith"
}'

And several order_item "children" documents that you index using their parent ID:

curl -XPUT localhost:9200/order_items/order_item/1?parent=82 -d '{
     "ID": 1,
     "Name": "Shoes",
     "Price": 9.99
}'
curl -XPUT localhost:9200/order_items/order_item/2?parent=82 -d '{
     "ID": 2,
     "Name": "Hat",
     "Price": 19.99
}'

In the second case, your order document would contain all order items in a nested OrderItems property and would look like this:

curl -XPUT localhost:9200/orders/order/82 -d '{
    "ID": 82,
    "Customer": "John Smith",
    "OrderItems": [
      {
        "ID": 1,
        "Name": "Shoes",
        "Price": 9.99
      },{
        "ID": 2,
        "Name": "Hat",
        "Price": 19.99
      }
    ]
}'

与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
OGeek|极客中国-欢迎来到极客的世界,一个免费开放的程序员编程交流平台!开放,进步,分享!让技术改变生活,让极客改变未来! Welcome to OGeek Q&A Community for programmer and developer-Open, Learning and Share
Click Here to Ask a Question

...