GCP FireStore：一对多嵌套场景下Map of Maps vs Array of Maps选型咨询

Great question! Let's break down the tradeoffs between these two approaches for your Account → Contact → Address hierarchical structure (all one-to-many relationships stored in a single document, no subcollections/subdocuments).

First, let's recap the two structures for clarity:

Map of Maps Example (your preferred approach) #

{
  "Name": "My Account Name",
  "Contact": {
    "Contact123": {
      "FirstName": "F1",
      "LastName": "L1",
      "Address": {
        "Address1231": {
          "Street": "Street1",
          "City": "City1"
        },
        "Address1232": {
          "Street": "Street2",
          "City": "City2"
        }
      }
    },
    "Contact234": {
      "FirstName": "F2",
      "LastName": "L2",
      "Address": {
        "Address2341": {
          "Street": "Street3",
          "City": "City3"
        },
        "Address2342": {
          "Street": "Street4",
          "City": "City4"
        }
      }
    }
  }
}

Array of Maps Example #

{
  "Name": "My Account Name",
  "Contact": [
    {
      "ContactId": "Contact123",
      "FirstName": "F1",
      "LastName": "L1",
      "Address": [
        {
          "AddressId": "Address1231",
          "Street": "Street1",
          "City": "City1"
        },
        {
          "AddressId": "Address1232",
          "Street": "Street2",
          "City": "City2"
        }
      ]
    },
    {
      "ContactId": "Contact234",
      "FirstName": "F2",
      "LastName": "L2",
      "Address": [
        {
          "AddressId": "Address2341",
          "Street": "Street3",
          "City": "City3"
        },
        {
          "AddressId": "Address2342",
          "Street": "Street4",
          "City": "City4"
        }
      ]
    }
  ]
}

Map of Maps: Pros & Cons #

Pros #

• Direct, O(1) access to child entities: You can instantly fetch a specific Contact or Address using its ID (e.g., Account.Contact.Contact123 or Account.Contact.Contact123.Address.Address1231) without looping through a list. This is a huge win for scripts or code that needs frequent targeted updates/deletes.
• Built-in ID uniqueness: Map keys are inherently unique, so you don't have to write extra business logic to prevent duplicate Contact/Address IDs from being added.
• Simplified updates/deletes: Removing or modifying a specific child entity is straightforward—just target the key directly, no need to first search the collection for the matching ID.

Cons #

• No inherent order: Most document database map/object types are unordered. If your business requires retaining the order of Contacts or Addresses (e.g., creation order), you'll need to maintain a separate array of IDs to track sequence, adding complexity.
• Clunky range queries: Filtering for all Addresses in a specific city, or counting Contacts with a certain last name, requires iterating through all map keys. This is less efficient and more code-heavy than using array-specific query operators (like MongoDB's $elemMatch).
• Less intuitive structure: For developers unfamiliar with this pattern, nested maps can be harder to visualize and debug compared to standard arrays of objects.

Array of Maps: Pros & Cons #

Pros #

• Native order preservation: Arrays maintain the order of elements, so you don't need extra work to track the sequence of Contacts/Addresses (critical for use cases like displaying recent contacts first).
• Better query support: Most document databases have robust built-in tools for querying, filtering, and aggregating array data. You can easily write queries to find all matching child entities without manual iteration.
• More intuitive for teams: Arrays of objects align with common programming patterns, making the codebase easier to understand, maintain, and onboard new developers to.

Cons #

• O(n) access time: To find a specific Contact or Address by ID, you have to loop through the entire array. This becomes slower and more cumbersome as the number of child entities grows.
• No built-in ID uniqueness: You'll need to add custom validation logic to prevent duplicate IDs from being inserted into the array, increasing development overhead.
• Complex updates/deletes: Modifying or removing a specific child entity requires first searching the array to find the matching ID, then updating that position—adding extra steps to your code.

When to Choose Which? #

Map of Maps is the better choice if:

• Your primary operations are targeted lookups, updates, or deletes using entity IDs.
• You don't need to preserve the order of child entities.
• Performance for direct ID-based access is a top priority.

Array of Maps is the better choice if:

• You need to maintain the order of Contacts/Addresses.
• You frequently run range queries, filters, or aggregations on child entities.
• Team readability and standardization are more important than micro-optimizations for ID-based access.

内容的提问来源于stack exchange，提问作者Ayush