在Redshift

Question

有没有办法从一个S3 JSON文件加载多个表Redshift并忽略那些与我的JSON路径不完全匹配的对象？

我在Redshift中有以下两个表：

create table employee
(
       employee_id int,
       employee_name varchar(50),
       employee_age int,
       department_id int,
       created_date timestamp
);
create table department
(
       department_id int,
       department_name varchar(50),
       created_date timestamp
);

我的JSON文件＆＃34; employee_department.json＆＃34;（包含员工和部门数据）：

{
    "employee_id": 1,
    "employee_name": "Jack",
    "employee_age ": 38,
    "department_id": 100,
    "created_date": "2016-11-07 10:00:00"
}
{
    "employee_id": 2,
    "employee_name": "Jill",
    "employee_age ": 26,
    "department_id": 101,
    "created_date": "2016-11-07 11:00:00"
}
{
    "department_id ": 1,
    "department_name ": "Sales",
    "created_date": "2016-01-01 03:00:00"
}
{
    "department_id ": 2,
    "department_name ": "Finance",
    "created_date": "2016-01-01 04:30:00"
}

清单文件＆＃34; employee_department_manifest.manifest＆＃34;：

{
  "entries": [
    {"url":"s3://mybucket/employee_department.json","mandatory":true}
  ]
}

两个JSON路径文件＆＃34; employee_jsonpath.json＆＃34; ＆安培; ＆＃34; department_jsonpath.json＆＃34;：

{
    "jsonpaths": [
        "$['employee_id']",
        "$['employee_name']",
        "$['employee_age']",
        "$['department_id']",
        "$['created_date']"
    ]
}


{
    "jsonpaths": [
        "$['department_id']",
        "$['department_name']",
        "$['created_date']"
    ]
}

以下COPY命令可以使用，但是根据下面的屏幕截图，我最终在每个表中有4条记录而不是预期的2条，并且错误的department_id数据被加载到departments表中。有没有办法将COPY限制为完全符合JSON路径文件的记录，而忽略其他文件？我意识到在这种情况下我可以将JSON分成两个单独的文件，但是知道这是否可能会很好。

COPY employee
FROM 's3://mybucket/employee_department_manifest.manifest'
CREDENTIALS ''
JSON 's3://mybucket/employee_jsonpath.json'
MANIFEST;
COPY department
FROM 's3://mybucket/employee_department_manifest.manifest'
CREDENTIALS ''
JSON 's3://mybucket/department_jsonpath.json'
MANIFEST;

select * from employee;

select * from department;