如何解析嵌套的records JSON数组？附示例JSON结构

Great question! Parsing nested JSON arrays with duplicate keys like records is super common—especially with APIs like Salesforce's, which your example clearly comes from. Let's break down how to handle this step by step, with practical examples in popular languages.

General Approach #

The core idea is to traverse the JSON structure layer by layer:

1. Parse the top-level JSON object to access the initial records array.
2. Iterate over each item in the top-level records.
3. For each item, check if it contains a field (like OrderItems here) that holds another object with its own records array.
4. If nested records exist, repeat the process—use recursion if you need to handle unknown or deep levels of nesting.

Example Implementations #

Python (Using Built-in json Module) #

Python makes this straightforward with its native JSON tools:

import json

# Your full JSON data (replace with your actual input string/file)
json_str = """
{
    "totalSize":17,
    "done":true,
    "records":[
        {
            "attributes":{
                "type":"Order",
                "url":"/services/data/v51.0/sobjects/Order/8011m0000008c8WAAQ"
            },
            "Id":"8011m0000008c8WAAQ",
            "DocumentType__c":null,
            "Orderno__c":null,
            "BiitoCustomerNo__c":null,
            "SelltoCustomerNo__c":null,
            "OrderDate__c":null,
            "PostingDate__c":null,
            "DocumentDate__c":null,
            "LocationCode__c":null,
            "Narration__c":null,
            "OrderCode__c":"Order - 000013",
            "OrderItems":null
        },
        {
            "Id":"8011m0000008c0SAAQ",
            "DocumentType__c":null,
            "Orderno__c":null,
            "BiitoCustomerNo__c":null,
            "SelltoCustomerNo__c":null,
            "OrderDate__c":null,
            "PostingDate__c":null,
            "DocumentDate__c":null,
            "LocationCode__c":null,
            "Narration__c":null,
            "OrderCode__c":"Order - 000008",
            "OrderItems":{
                "totalSize":2,
                "done":true,
                "records":[
                    {
                        "attributes":{
                            "type":"OrderItem",
                            "url":"/services/data/v51.0/sobjects/OrderItem/8021m000000DBirAAG"
                        },
                        "Id":"8021m000000DBirAAG",
                        "Document_Type__c":null,
                        "Sell_to_Customer_No__c":null,
                        "Item_Code__c":"OP-000004",
                        "Variant_Code__c":null,
                        "Location_Code__c":null,
                        "Quantity":2,
                        "Sales_Quantity__c":null,
                        "Shipment_Date__c":null
                    },
                    {
                        "attributes":{
                            "type":"OrderItem",
                            "url":"/services/data/v51.0/sobjects/OrderItem/8021m000000DBiqAAG"
                        },
                        "Id":"8021m000000DBiqAAG",
                        "Document_Type__c":null,
                        "Sell_to_Customer_No__c":null,
                        "Item_Code__c":"OP-000003",
                        "Variant_Code__c":null,
                        "Location_Code__c":null,
                        "Quantity":2,
                        "Sales_Quantity__c":null,
                        "Shipment_Date__c":null
                    }
                ]
            }
        }
    ]
}
"""

# Parse the top-level JSON
data = json.loads(json_str)

# Iterate over top-level orders
for order in data["records"]:
    print(f"Order: {order['OrderCode__c']} (ID: {order['Id']})")
    
    # Check for nested OrderItems records
    if order.get("OrderItems") and order["OrderItems"].get("records"):
        print("  Line Items:")
        for item in order["OrderItems"]["records"]:
            print(f"    - {item['Item_Code__c']}: Quantity {item['Quantity']} (ID: {item['Id']})")

JavaScript (Using JSON.parse) #

For frontend or Node.js environments, here's how to handle it:

// Your full JSON string
const jsonStr = `{
    "totalSize":17,
    "done":true,
    "records":[
        {
            "attributes":{
                "type":"Order",
                "url":"/services/data/v51.0/sobjects/Order/8011m0000008c8WAAQ"
            },
            "Id":"8011m0000008c8WAAQ",
            "DocumentType__c":null,
            "Orderno__c":null,
            "BiitoCustomerNo__c":null,
            "SelltoCustomerNo__c":null,
            "OrderDate__c":null,
            "PostingDate__c":null,
            "DocumentDate__c":null,
            "LocationCode__c":null,
            "Narration__c":null,
            "OrderCode__c":"Order - 000013",
            "OrderItems":null
        },
        {
            "Id":"8011m0000008c0SAAQ",
            "DocumentType__c":null,
            "Orderno__c":null,
            "BiitoCustomerNo__c":null,
            "SelltoCustomerNo__c":null,
            "OrderDate__c":null,
            "PostingDate__c":null,
            "DocumentDate__c":null,
            "LocationCode__c":null,
            "Narration__c":null,
            "OrderCode__c":"Order - 000008",
            "OrderItems":{
                "totalSize":2,
                "done":true,
                "records":[
                    {
                        "attributes":{
                            "type":"OrderItem",
                            "url":"/services/data/v51.0/sobjects/OrderItem/8021m000000DBirAAG"
                        },
                        "Id":"8021m000000DBirAAG",
                        "Document_Type__c":null,
                        "Sell_to_Customer_No__c":null,
                        "Item_Code__c":"OP-000004",
                        "Variant_Code__c":null,
                        "Location_Code__c":null,
                        "Quantity":2,
                        "Sales_Quantity__c":null,
                        "Shipment_Date__c":null
                    },
                    {
                        "attributes":{
                            "type":"OrderItem",
                            "url":"/services/data/v51.0/sobjects/OrderItem/8021m000000DBiqAAG"
                        },
                        "Id":"8021m000000DBiqAAG",
                        "Document_Type__c":null,
                        "Sell_to_Customer_No__c":null,
                        "Item_Code__c":"OP-000003",
                        "Variant_Code__c":null,
                        "Location_Code__c":null,
                        "Quantity":2,
                        "Sales_Quantity__c":null,
                        "Shipment_Date__c":null
                    }
                ]
            }
        }
    ]
}`;

// Parse the JSON
const data = JSON.parse(jsonStr);

// Loop through orders
data.records.forEach(order => {
    console.log(`Order: ${order.OrderCode__c} (ID: ${order.Id})`);
    
    // Check for nested line items
    if (order.OrderItems && order.OrderItems.records) {
        console.log("  Line Items:");
        order.OrderItems.records.forEach(item => {
            console.log(`    - ${item.Item_Code__c}: Quantity ${item.Quantity} (ID: ${item.Id})`);
        });
    }
});

Handling Deep/Unknown Nesting #

If your JSON has multiple levels of nested records (not just one), use a recursive function to automatically traverse all levels. Here's a Python example:

def traverse_records(records, depth=0):
    indent = "  " * depth
    for item in records:
        # Print a basic identifier (adjust based on your data's fields)
        if "Id" in item:
            print(f"{indent}ID: {item['Id']}")
        # Check all fields for nested records
        for key, value in item.items():
            if isinstance(value, dict) and "records" in value:
                print(f"{indent}Nested records found under '{key}':")
                traverse_records(value["records"], depth + 1)

# Call the recursive function on the top-level records
traverse_records(data["records"])

This function will dig into every nested records array, no matter how deep it goes.

内容的提问来源于stack exchange，提问作者pandix