我正在尝试使用Newtonsoft.Json.Net35版本4.0.2.0来反序列化包含空值的ADO.NET DataTable。序列化工作正常:
[Test]
public void SerializeDataTableWithNull()
{
var table = new DataTable();
table.Columns.Add("item");
table.Columns.Add("price", typeof(double));
table.Rows.Add("shirt", 49.99);
table.Rows.Add("pants", 54.99);
table.Rows.Add("shoes"); // no price
var json = JsonConvert.SerializeObject(table);
Assert.AreEqual(@"["
+ @"{""item"":""shirt"",""price"":49.99},"
+ @"{""item"":""pants"",""price"":54.99},"
+ @"{""item"":""shoes"",""price"":null}]", json);
}
如果缺少值,反序列化可以正常工作:
[Test]
public void DerializeDataTableWithImplicitNull()
{
const string json = @"["
+ @"{""item"":""shirt"",""price"":49.99},"
+ @"{""item"":""pants"",""price"":54.99},"
+ @"{""item"":""shoes""}]";
var table = JsonConvert.DeserializeObject<DataTable>(json);
Assert.AreEqual("shirt", table.Rows[0]["item"]);
Assert.AreEqual("pants", table.Rows[1]["item"]);
Assert.AreEqual("shoes", table.Rows[2]["item"]);
Assert.AreEqual(49.99, (double)table.Rows[0]["price"], 0.01);
Assert.AreEqual(54.99, (double)table.Rows[1]["price"], 0.01);
Assert.IsInstanceOf(typeof(System.DBNull), table.Rows[2]["price"]);
}
但是,如果值显式为null:
[Test]
public void DerializeDataTableWithExplicitNull()
{
const string json = @"["
+ @"{""item"":""shirt"",""price"":49.99},"
+ @"{""item"":""pants"",""price"":54.99},"
+ @"{""item"":""shoes"",""price"":null}]";
var table = JsonConvert.DeserializeObject<DataTable>(json);
Assert.AreEqual("shirt", table.Rows[0]["item"]);
Assert.AreEqual("pants", table.Rows[1]["item"]);
Assert.AreEqual("shoes", table.Rows[2]["item"]);
Assert.AreEqual(49.99, (double)table.Rows[0]["price"], 0.01);
Assert.AreEqual(54.99, (double)table.Rows[1]["price"], 0.01);
Assert.IsInstanceOf(typeof(System.DBNull), table.Rows[2]["price"]);
}
DeserializeObject抛出“System.ArgumentException:无法将Column'price'设置为null。请改用DBNull。”以下解决方法适用于我的特定JSON:
var regex = new Regex(@",?""[_\w]+"":null");
var nullless = regex.Replace(json, string.Empty);
var table = JsonConvert.DeserializeObject<DataTable>(nullless);
但与所有基于正则表达式的kludges一样,这显然是脆弱的。
最后,问题:
提前致谢,
谢
答案 0 :(得分:0)
看起来这是一个可以通过替换
轻松修复的错误dr[columnName] = reader.Value
带
dr[columnName] = reader.Value ?? System.DBNull.Value
Newtonsoft.Json.Converters.DataTableConverter中的。我在跟踪器中输入了issue。