如何從 C# 中的 json 創建數據表（visual basic）

Question

我繼承了一個使用 Visual Studio 的項目（我認為是 .NET Framework 4.7.2）。 我收到來自 API 端點的響應，其中包含多個條目，如下所示：

[
    {
        "transaction_id": "00000352",
        "transaction_type": "New",
        "transaction_date": "2018-08-23T00:00:00",
        "sold_to_id": "00026",
        "customer_po_number": "34567",
        "po_date": "2018-08-23T00:00:00",
        "notes": null,
        "batch_code": "######",
        "location_id": "MN",
        "req_ship_date": "2018-08-28T00:00:00",
        "fiscal_period": 8,
        "fiscal_year": 2018,
        "currency_id": "USD",
        "original_invoice_number": null,
        "bill_to_id": "00026",
        "customer_level": null,
        "terms_code": "30-2",
        "distribution_code": "01",
        "invoice_number": null,
        "invoice_date": "2018-08-23T00:00:00",
        "sales_rep_id_1": null,
        "sales_rep_id_1_percent": 100,
        "sales_rep_id_1_rate": 0,
        "sales_rep_id_2_id": "SA",
        "sales_rep_id_2_percent": 100,
        "sales_rep_id_2_rate": 10,
        "ship_to_id": null,
        "ship_via": null,
        "ship_method": null,
        "ship_number": null,
        "ship_to_name": "",
        "ship_to_attention": null,
        "ship_to_address_1": null,
        "ship_to_address_2": null,
        "ship_to_city": null,
        "ship_to_region": null,
        "ship_to_country": "USA",
        "ship_to_postal_code": null,
        "actual_ship_date": null,
        "order_line_list": [
            {
                "entry_number": 1,
                "item_id": "5' Glass/Wood Table",
                "description": "Glass/ Wood Combo Coffee Table",
                "customer_part_no": null,
                "additional_description": null,
                "location_id": "MN",
                "quantity_ordered": 11,
                "units": "EA",
                "quantity_shipped": 0,
                "quantity_backordered": 0,
                "req_ship_date": null,
                "unit_price": 599.99,
                "extended_price": 6599.89,
                "reason_code": null,
                "account_code": null,
                "inventory_gl_account": "120000",
                "sales_gl_account": "400000",
                "cogs_gl_account": "500000",
                "sales_category": null,
                "tax_class": 0,
                "promo_id": null,
                "price_id": null,
                "discount_type": 0,
                "discount_percentage": 0,
                "discount_amount": 0,
                "sales_rep_id_1": null,
                "sales_rep1_percent": null,
                "sales_rep1_rate": null,
                "sales_rep_id_2": "SA",
                "sales_rep2_percent": 100,
                "sales_rep2_rate": 5,
                "unit_cost": 241.4467,
                "extended_cost": 2655.91,
                "status": "Open",
                "extended_list": [],
                "serial_list": []
            }
        ],
        "tax_group_id": "ATL",
        "taxable": false,
        "tax_class_adjustment": 0,
        "freight": 0,
        "tax_class_freight": 0,
        "tax_location_adjustment": null,
        "misc": 0,
        "tax_class_misc": 0,
        "sales_tax": 0,
        "taxable_sales": 0,
        "non_taxable_sales": 6599.89,
        "payment_list": []
    }
]

我想從這個響應中 select 單個鍵值對並創建一個數據表。 我對 C# 非常陌生（並且有些過敏），但我嘗試過這樣做：

var doc = JsonDocument.Parse(response.Content);
JsonElement root = doc.RootElement;

//Console.WriteLine(doc.RootElement);
Console.WriteLine(String.Format("Root Elem: {0}", doc.RootElement));

//creat data table
var jsonObject = JObject.Parse(response.Content);
Console.WriteLine(String.Format("jsonObject col: {0}", jsonObject));
DataTable jsdt = jsonObject[doc.RootElement].ToObject<DataTable>();

int totalRows = jsdt.Rows.Count;


Console.WriteLine(String.Format("jsonObject col: {0}", jsonObject));

Console.WriteLine(String.Format("totalRows col: {0}", totalRows));
Console.WriteLine(String.Format("jsdt col: {0}", jsdt.Rows.Count));

我似乎無法讓 DataTable 登錄到控制台，也不知道我是否真的在創建 DataTable。 我可以遍歷 JSON 響應以查找單個鍵值對，但是將內容寫入 DataTable 是我遇到的問題。 這就是我循環通過 JSON 的方式：

var users = root.EnumerateArray();

while (users.MoveNext())
{
    var user = users.Current;
    //System.Console.WriteLine(user);

    var props = user.EnumerateObject();

    while (props.MoveNext())
    {
        var prop = props.Current;
        //Console.WriteLine($"{prop.Name}: {prop.Value}");
        if (prop.Name == "transaction_date") {
            //Console.WriteLine(String.Format("key: {0}", $"{prop.Name}: {prop.Value}"));
        }
    }
}

go 如何以這種方式創建DataTable ？

更新：

我的最終目的是創建一個 CSV 並將其發送到另一台服務器。

我認識到 CSV 文件基本上是原始值的二維數組，但我的 JSON 對象包含嵌套的 arrays 的復雜對象（例如[*].order_line_list[*] 為此，扁平化為多行就可以了，例如父名稱、父信息、嵌套項信息 1； 父名稱、父信息、嵌套項信息 2；

Answer 1

使用正則表達式：

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Text.RegularExpressions;
using System.IO;

namespace ConsoleApplication37
{
    class Program
    {
        const string FILENAME = @"c:\temp\test.txt";
        static void Main(string[] args)
        {
            string input = File.ReadAllText(FILENAME);
            string pattern = "\"(?'key'[^\"]+)\":\\s+(?'value'[^,]+)";
            MatchCollection matches = Regex.Matches(input, pattern, RegexOptions.Multiline);

            Dictionary<string, string> dict = matches.Cast<Match>()
                .GroupBy(x => x.Groups["key"].Value, y => y.Groups["value"].Value.Trim())
                .ToDictionary(x => x.Key, y => y.FirstOrDefault());


        }
    }
}

Answer 2

Your basic problem here is that you have a JSON array containing JSON objects with possibly non-primitive object and array values, and you would like to eventually export that to CSV. 但是，CSV 文件本質上是原始值的二維數組，因此您需要將非原始值展平為原始值。 為此，您可以

1. 通過為嵌套值添加額外的列來水平擴展 CSV。
2. 通過為嵌套值添加額外的行來垂直擴展 CSV。

您選擇了選項 2，多行即可，例如父名稱、父信息、嵌套項信息 1； 父名稱、父信息、嵌套項信息 2；

To do that, we're going to need two third-party tools, the LINQ to JSON from Json.NET to parse your arbitrary JSON objects, and CsvWriter from CsvHelper to write a CSV file.

首先，引入以下擴展方法，將 LINQ 到 JSON JObject展平為原始 Z0ECD11C1D7A287401D148A23BBD7A2F 的可枚舉鍵/值對的可枚舉：

public static partial class JsonExtensions
{
    public static IEnumerable<IEnumerable<KeyValuePair<string, JValue>>> FlattenChildrenIntoRows(this JObject obj)
    {
        bool anyReturned = false;
        var simpleChildren = obj.Properties().Where(p => p.Value is JValue).Select(p => new KeyValuePair<string, JValue>(p.Name, (JValue)p.Value)).ToList();
        var complexChildren = obj.Properties().Where(p => !(p.Value is JValue)).Select(p => new KeyValuePair<string, JToken>(p.Name, p.Value));
        foreach (var property in complexChildren)
        {
            foreach (var flattenedItems in FlattenChildrenIntoRows(property))
            {
                yield return simpleChildren.Concat(flattenedItems);
                anyReturned = true;
            }
        }

        if (!anyReturned)
            yield return simpleChildren;
    }

    static IEnumerable<IEnumerable<KeyValuePair<string, JValue>>> FlattenChildrenIntoRows(KeyValuePair<string, JToken> property) 
    {
        if (property.Value is JValue value)
        {
            yield return new[] { new KeyValuePair<string, JValue>(property.Key, value) };
        }
        else if (property.Value is JArray array)
        {
            foreach (var item in array)
                foreach (var flattenedItem in FlattenChildrenIntoRows(new KeyValuePair<string, JToken>(property.Key, item)))
                    yield return flattenedItem;
        }
        else if (property.Value is JObject obj)
        {
            foreach (var flattenedItems in FlattenChildrenIntoRows(obj))
                yield return flattenedItems.Select(p => new KeyValuePair<string, JValue>(property.Key + "." + p.Key, p.Value));
        }
        else // JRaw, JContainer
        {
            throw new NotImplementedException();
        }
    }
}

接下來，引入以下方法將Dictionary<string, string>對象列表寫入 CSV 文件，使字典鍵為 CSV 列名：

public static partial class CsvExtensions
{
    public static void WriteCsv(TextWriter writer, IReadOnlyList<IDictionary<string, string>> objects, CsvConfiguration config)
    {
        var headers = objects.SelectMany(o => o.Keys).Distinct();
        using (var csv = new CsvWriter(writer, config))
        {
            foreach (var header in headers)
                csv.WriteField(header);
            csv.NextRecord();
            foreach (var obj in objects)
            {
                foreach (var header in headers)
                    if (obj.TryGetValue(header, out var value))
                        csv.WriteField(value);
                    else
                        csv.WriteField(null);
                csv.NextRecord();
            }
        }           
    }
}

現在我們可以結合這兩種方法來讀取 JSON 數組，將其展平，然后寫入 CSV，如下所示：

public static partial class CsvExtensions
{
    /// Convert a JSON file to a CSV file
    public static void ConvertJsonToCsv(string inputPath, string outputPath)
    {
        using var reader = new StreamReader(inputPath);
        using var writer = new StreamWriter(outputPath);
        ConvertJsonToCsv(reader, writer);
    }
    
    /// Read JSON from the incoming TextReader, convert to CSV and write to the incoming TextWriter
    public static void ConvertJsonToCsv(TextReader reader, TextWriter writer)
    {
        // Disable reformatting of dates
        var settings = new JsonSerializerSettings
        {
            DateParseHandling = DateParseHandling.None,
        };
        
        using var jsonReader = new JsonTextReader(reader) { CloseInput = false };
        
        // Deserialize to an array of objects
        
        var array = JsonSerializer.CreateDefault(settings).Deserialize<List<JObject>>(jsonReader);
        
        // Flatten and convert to a List<Dictionary<string, string>>
        
        var list = array
            // Flatten each child into an enumerable of enumerable of primitive key/value pairs
            .Select(o => o.FlattenChildrenIntoRows())
            // Project and flatten to an overall enumerable of enumerable of pairs
            .SelectMany(pairs => pairs)
            // Convert to an overall enumerable of dictionaries
            .Select(pairs => pairs.ToDictionary(p => p.Key, p => p.Value.Value == null ? null : (string)Convert.ChangeType(p.Value.Value, typeof(string), CultureInfo.InvariantCulture)))
            // Materialize as a list of dictionaries
            .ToList();
        
        // Write to CSV
        var config = new CsvConfiguration(CultureInfo.InvariantCulture);
        CsvExtensions.WriteCsv(writer, list, config);
    }
}

您可能會這樣稱呼：

using var reader = new StringReader(response.Content);
using var writer = new StringWriter();

CsvExtensions.ConvertJsonToCsv(reader, writer);

var csvText = writer.ToString();

Console.WriteLine(csvText);

演示小提琴在這里。