如何在 kusto 查询语言中解析 json 数组答案

【问题标题】：How to parse json array in kusto query language如何在 kusto 查询语言中解析 json 数组
【发布时间】：2021-03-10 23:36:01
【问题描述】：

如何在 kusto 查询语言中解析 json 数组。我有一个输出列，它具有 JSON 数组格式的值，如下所示。我需要解析它以获取两列形式的值。

{"count": 14
"value": [
        {
            "Total_Record_Count": 16608,
            "date": "2021-03-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 27254,
            "date": "2021-02-24T00:00:00Z"
        },
        {
            "Total_Record_Count": 6,
            "date": "2021-02-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 26964,
            "date": "2021-01-15T00:00:00Z"
        },
        {
            "Total_Record_Count": 134516,
            "date": "2020-12-18T00:00:00Z"
        },
        {
            "Total_Record_Count": 27345,
            "date": "2020-12-16T00:00:00Z"
        },
        {
            "Total_Record_Count": 521,
            "date": "2020-12-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 4,
            "date": "2020-11-02T00:00:00Z"
        },
        {
            "Total_Record_Count": 6,
            "date": "2020-10-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 1,
            "date": "2020-09-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 3,
            "date": "2020-08-03T00:00:00Z"
        },
        {
            "Total_Record_Count": 18,
            "date": "2020-07-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 18754,
            "date": "2020-06-16T00:00:00Z"
        },
        {
            "Total_Record_Count": 4451898,
            "date": "2020-06-08T00:00:00Z"
        }
    ]}

如何使用输出列名而不是使用完整的 json 数组来实现它。

【问题讨论】：

标签： azure-log-analytics kql

【解决方案1】：

请参阅下面的示例，该示例使用 mv-expand 运算符将数组分成行。 https://docs.microsoft.com/en-us/azure/data-explorer/kusto/query/mvexpandoperator

print d = dynamic(
{"count": 14,
"value": [
        {
            "Total_Record_Count": 16608,
            "date": "2021-03-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 27254,
            "date": "2021-02-24T00:00:00Z"
        },
        {
            "Total_Record_Count": 6,
            "date": "2021-02-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 26964,
            "date": "2021-01-15T00:00:00Z"
        },
        {
            "Total_Record_Count": 134516,
            "date": "2020-12-18T00:00:00Z"
        },
        {
            "Total_Record_Count": 27345,
            "date": "2020-12-16T00:00:00Z"
        },
        {
            "Total_Record_Count": 521,
            "date": "2020-12-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 4,
            "date": "2020-11-02T00:00:00Z"
        },
        {
            "Total_Record_Count": 6,
            "date": "2020-10-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 1,
            "date": "2020-09-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 3,
            "date": "2020-08-03T00:00:00Z"
        },
        {
            "Total_Record_Count": 18,
            "date": "2020-07-01T00:00:00Z"
        },
        {
            "Total_Record_Count": 18754,
            "date": "2020-06-16T00:00:00Z"
        },
        {
            "Total_Record_Count": 4451898,
            "date": "2020-06-08T00:00:00Z"
        }
    ]})
| project Value = d.['value'] 
| mv-expand Value
| project Count = tolong(Value.['Total_Record_Count']), Date = todatetime(Value.['date'])

Count	Date
4451898	2020-06-08 00:00:00.0000000
18754	2020-06-16 00:00:00.0000000
18	2020-07-01 00:00:00.0000000
3	2020-08-03 00:00:00.0000000
1	2020-09-01 00:00:00.0000000
6	2020-10-01 00:00:00.0000000
4	2020-11-02 00:00:00.0000000
521	2020-12-01 00:00:00.0000000
27345	2020-12-16 00:00:00.0000000
134516	2020-12-18 00:00:00.0000000
26964	2021-01-15 00:00:00.0000000
6	2021-02-01 00:00:00.0000000
27254	2021-02-24 00:00:00.0000000
16608	2021-03-01 00:00:00.0000000

【讨论】：

谢谢亚历山大。当我在 print d = dynamic（在此处粘贴 json 数组）中使用完整的 json 数组时，它的工作原理。但是我正在尝试使用列名来获取此 JSON 数组值作为输出。我找不到这样做的方法。
它应该工作相同。如果列类型是字符串而不是动态 - 您可以使用 parse_json(ColumnName) 函数将其转换为动态列
如果我们想对所有的 'Total_Record_Count' 求和，如何修改查询