雪花表中json数据的解析字段将多行插入到新的雪花表中答案

【问题标题】：Parse field of json data in Snowflake table insert multiple rows into new Snowflake table雪花表中json数据的解析字段将多行插入到新的雪花表中
【发布时间】：2021-08-17 15:57:47
【问题描述】：

我在雪花表“tbl”中有一个字段/列“反应”。该表包含几列数据和许多记录。 'reactions' 字段是一个 json 数组。 json 数据可能有多个反应对象（由“名称”表示）并列出了具有反应的“用户”（参见下面的示例数组）。我想将此数据字段解析到一个新表中，以便反应的每个用户 ID 都是它自己的记录，并包含反应“名称”以及源表“tbl”中的一些其他数据列。我试图展平数据并使用 json 解析，但我未能成功达到所需的输出。非常感谢您提供的任何支持或指导！

包含 json 数组的“反应”字段值示例：

[{
    "name": "wave",
    "users": ["UM6G1DGBG", "UUW2KMQF8", "UG47U9K9N", "U01KCT0KJHL", "U7GRE4E9M", "UU1V5U3B4", "U01S7RRG21Y", "U013J1D5NFQ", "U020AS49MFA", "U010F5EDDQV", "U01ETTJHY4F"],
    "count": 11
}, {
    "name": "wave::skin-tone-3",
    "users": ["UUBFUS220"],
    "count": 1
}, {
    "name": "fire",
    "users": ["U01ETTJHY4F"],
    "count": 1
}]

所需的输出将如下所示：

OtherData1	UserID	ReactionName
Attribute1	UM6G1DGBG	wave
Attribute1	UUW2KMQF8	wave
Attribute1	UG47U9K9N	wave
...	...	...
Attribute1	U01ETTJHY4F	fire

【问题讨论】：

标签： json parsing snowflake-cloud-data-platform

【解决方案1】：

您可以使用以下方法创建您想要的结果

with tbl as (select 'foo' other_data, parse_json('[{
    "name": "wave",
    "users": ["UM6G1DGBG", "UUW2KMQF8", "UG47U9K9N", "U01KCT0KJHL", "U7GRE4E9M", "UU1V5U3B4", "U01S7RRG21Y", "U013J1D5NFQ", "U020AS49MFA", "U010F5EDDQV", "U01ETTJHY4F"],
    "count": 11
}, {
    "name": "wave::skin-tone-3",
    "users": ["UUBFUS220"],
    "count": 1
}, {
    "name": "fire",
    "users": ["U01ETTJHY4F"],
    "count": 1
}]') users)

select 
    other_data,  
    user_ids.value,  
    u.value:name,  
    u.value  
from tbl,  
lateral flatten(users) u,  
lateral flatten(u.value:users) user_ids
;



OTHER_DATA  VALUE   U.VALUE:NAME    VALUE
foo "UM6G1DGBG" "wave"  {    "count": 11,    "name": "wave",    "users": [      "UM6G1DGBG",      "UUW2KMQF8",      "UG47U9K9N",      "U01KCT0KJHL",      "U7GRE4E9M",      "UU1V5U3B4",      "U01S7RRG21Y",      "U013J1D5NFQ",      "U020AS49MFA",      "U010F5EDDQV",      "U01ETTJHY4F"    ]  }
foo "UUW2KMQF8" "wave"  {    "count": 11,    "name": "wave",    "users": [      "UM6G1DGBG",      "UUW2KMQF8",      "UG47U9K9N",      "U01KCT0KJHL",      "U7GRE4E9M",      "UU1V5U3B4",      "U01S7RRG21Y",      "U013J1D5NFQ",      "U020AS49MFA",      "U010F5EDDQV",      "U01ETTJHY4F"    ]  }
foo "UG47U9K9N" "wave"  {    "count": 11,    "name": "wave",    "users": [      "UM6G1DGBG",      "UUW2KMQF8",      "UG47U9K9N",      "U01KCT0KJHL",      "U7GRE4E9M",      "UU1V5U3B4",      "U01S7RRG21Y",      "U013J1D5NFQ",      "U020AS49MFA",      "U010F5EDDQV",      "U01ETTJHY4F"    ]  }
foo "U01KCT0KJHL"   "wave"  {    "count": 11,    "name": "wave",    "users": [      "UM6G1DGBG",      "UUW2KMQF8",      "UG47U9K9N",      "U01KCT0KJHL",      "U7GRE4E9M",      "UU1V5U3B4",      "U01S7RRG21Y",      "U013J1D5NFQ",      "U020AS49MFA",      "U010F5EDDQV",      "U01ETTJHY4F"    ]  }
foo "U7GRE4E9M" "wave"  {    "count": 11,    "name": "wave",    "users": [      "UM6G1DGBG",      "UUW2KMQF8",      "UG47U9K9N",      "U01KCT0KJHL",      "U7GRE4E9M",      "UU1V5U3B4",      "U01S7RRG21Y",      "U013J1D5NFQ",      "U020AS49MFA",      "U010F5EDDQV",      "U01ETTJHY4F"    ]  }
foo "UU1V5U3B4" "wave"  {    "count": 11,    "name": "wave",    "users": [      "UM6G1DGBG",      "UUW2KMQF8",      "UG47U9K9N",      "U01KCT0KJHL",      "U7GRE4E9M",      "UU1V5U3B4",      "U01S7RRG21Y",      "U013J1D5NFQ",      "U020AS49MFA",      "U010F5EDDQV",      "U01ETTJHY4F"    ]  }
foo "U01S7RRG21Y"   "wave"  {    "count": 11,    "name": "wave",    "users": [      "UM6G1DGBG",      "UUW2KMQF8",      "UG47U9K9N",      "U01KCT0KJHL",      "U7GRE4E9M",      "UU1V5U3B4",      "U01S7RRG21Y",      "U013J1D5NFQ",      "U020AS49MFA",      "U010F5EDDQV",      "U01ETTJHY4F"    ]  }
foo "U013J1D5NFQ"   "wave"  {    "count": 11,    "name": "wave",    "users": [      "UM6G1DGBG",      "UUW2KMQF8",      "UG47U9K9N",      "U01KCT0KJHL",      "U7GRE4E9M",      "UU1V5U3B4",      "U01S7RRG21Y",      "U013J1D5NFQ",      "U020AS49MFA",      "U010F5EDDQV",      "U01ETTJHY4F"    ]  }
foo "U020AS49MFA"   "wave"  {    "count": 11,    "name": "wave",    "users": [      "UM6G1DGBG",      "UUW2KMQF8",      "UG47U9K9N",      "U01KCT0KJHL",      "U7GRE4E9M",      "UU1V5U3B4",      "U01S7RRG21Y",      "U013J1D5NFQ",      "U020AS49MFA",      "U010F5EDDQV",      "U01ETTJHY4F"    ]  }
foo "U010F5EDDQV"   "wave"  {    "count": 11,    "name": "wave",    "users": [      "UM6G1DGBG",      "UUW2KMQF8",      "UG47U9K9N",      "U01KCT0KJHL",      "U7GRE4E9M",      "UU1V5U3B4",      "U01S7RRG21Y",      "U013J1D5NFQ",      "U020AS49MFA",      "U010F5EDDQV",      "U01ETTJHY4F"    ]  }
foo "U01ETTJHY4F"   "wave"  {    "count": 11,    "name": "wave",    "users": [      "UM6G1DGBG",      "UUW2KMQF8",      "UG47U9K9N",      "U01KCT0KJHL",      "U7GRE4E9M",      "UU1V5U3B4",      "U01S7RRG21Y",      "U013J1D5NFQ",      "U020AS49MFA",      "U010F5EDDQV",      "U01ETTJHY4F"    ]  }
foo "UUBFUS220" "wave::skin-tone-3" {    "count": 1,    "name": "wave::skin-tone-3",    "users": [      "UUBFUS220"    ]  }
foo "U01ETTJHY4F"   "fire"  {    "count": 1,    "name": "fire",    "users": [      "U01ETTJHY4F"    ]  }```

So to create the table do something like `create table flattened_tbl(other_data, user_name, user_data) as select other_data, u.value:name, u.value from tbl, lateral flatten(users) u;`

【讨论】：

嗨@Nat Taylor 感谢您的回复。我相信这会让我得到每个名称的记录，而不是每个用户 ID 的记录（例如用户 ID 值“U01ETTJHY4F”）。我可能最初沟通得不够好，所以我用所需的输出表修改了帖子描述，以更好地传达我的目标。
我认为我编辑了答案以反映您的用例。