Transform and filter array of structs with parent struct field name.

huangapple go评论68阅读模式
英文:

Transform and filter array of structs with parent struct field name

问题

以下是您要翻译的内容:

I am trying to do one more step further than this StackOverflow post (https://stackoverflow.com/questions/74299990/convert-struct-of-structs-to-array-of-structs-pulling-struct-field-name-inside) where I need to pull the struct field name, filter each struct array based on a condition of role values and transform each struct element into a new struct with the extracted struct field name.

Input:

 |-- a: array (nullable = true)
 |    |-- element: struct (containsNull = true)
 |    |    |-- struct_key: string (nullable = true)
 |    |    |-- two: array (nullable = true)
 |    |    |    |-- element: struct (containsNull = true)
 |    |    |    |    |-- name: string (nullable = true)
 |    |    |    |    |-- role: string (nullable = true)
 |    |-- element: struct (containsNull = true)
 |    |    |-- struct_key: string (nullable = true)
 |    |    |-- two: array (nullable = true)
 |    |    |    |-- element: struct (containsNull = true)
 |    |    |    |    |-- name: string (nullable = true)
 |    |    |    |    |-- role: string (nullable = true)
{
	"a": [{
			"two": [{
				"name": "person1",
				"role": "role1"
			},
			{
				"name": "person2",
				"role": "role1"
			},
			{
				"name": "person3",
				"role": "role2"
			}],
			"struct_key": "test1"
		},
		{
			"two": [{
				"name": "person4",
				"role": "role1"
			},
			{
				"name": "person5",
				"role": "role1"
			},
			{
				"name": "person6",
				"role": "role2"
			}],
			"struct_key": "test2"
		}
	]
}
input ={
    "a": [{
            "two": [{
                    "name": "person1",
                    "role": "role1"
                },
                {
                    "name": "person2",
                    "role": "role1"
                },
                {
                    "name": "person3",
                    "role": "role2"
                }
            ],
            "struct_key": "test1"
        },
        {
            "two": [{
                    "name": "person4",
                    "role": "role1"
                },
                {
                    "name": "person5",
                    "role": "role1"
                },
                {
                    "name": "person6",
                    "role": "role2"
                }
            ],
            "struct_key": "test2"
        }
    ]
}

df = spark.read.json(sc.parallelize([input]))
print(df.selectExpr('inline(a)').schema)

Expected output after filtering (for roles) and new struct transformation:

 |-- role_output: array (nullable = true)
 |    |-- element: struct (containsNull = true)
 |    |    |-- struct_key: string (nullable = true)
 |    |    |-- name: string (nullable = true)
{
	"role1_output": [
		{
			"struct_key": "test1",
			"name": "person1"
		}, 
		{
			"struct_key": "test1",
			"name": "person2"
		},
		{
			"struct_key": "test2",
			"name": "person4"
		},
		{
			"struct_key": "test2",
			"name": "person5"
		}
	]
}

{
	"role2_output": [
		{
			"struct_key": "test1",
			"name": "person3"
		}, 
		{
			"struct_key": "test2",
			"name": "person6"
		}
	]
}

我已经为您提供了所需的翻译部分,没有包括其他内容。如果您需要更多帮助,请随时告诉我。

英文:

I am trying to do one more step further than this StackOverflow post (https://stackoverflow.com/questions/74299990/convert-struct-of-structs-to-array-of-structs-pulling-struct-field-name-inside) where I need to pull the struct field name, filter each struct array based on a condition of role values and transform each struct element into a new struct with the extracted struct field name.

Input:

 |-- a: array (nullable = true)
|    |-- element: struct (containsNull = true)
|    |    |-- struct_key: string (nullable = true)
|    |    |-- two: array (nullable = true)
|    |    |    |-- element: struct (containsNull = true)
|    |    |    |    |-- name: string (nullable = true)
|    |    |    |    |-- role: string (nullable = true)
|    |-- element: struct (containsNull = true)
|    |    |-- struct_key: string (nullable = true)
|    |    |-- two: array (nullable = true)
|    |    |    |-- element: struct (containsNull = true)
|    |    |    |    |-- name: string (nullable = true)
|    |    |    |    |-- role: string (nullable = true)
{
"a": [{
"two": [{
"name": "person1"
"role": "role1"
},
{
"name": "person2"
"role": "role1"
},
{
"name": "person3"
"role": "role2"
}],
"struct_key": "test1"
},
{
"two": [{
"name": "person4"
"role": "role1"
},
{
"name": "person5"
"role": "role1"
},
{
"name": "person6"
"role": "role2"
}],
"struct_key": "test2"
}
]
}
input ={
"a": [{
"two": [{
"name": "person1",
"role": "role1"
},
{
"name": "person2",
"role": "role1"
},
{
"name": "person3",
"role": "role2"
}
],
"struct_key": "test1"
},
{
"two": [{
"name": "person4",
"role": "role1"
},
{
"name": "person5",
"role": "role1"
},
{
"name": "person6",
"role": "role2"
}
],
"struct_key": "test2"
}
]
}
df = spark.read.json(sc.parallelize([input]))
print(df.selectExpr('inline(a)').schema)

Expected output after filtering (for roles) and new struct transformation:

 |-- role_output: array (nullable = true)
|    |-- element: struct (containsNull = true)
|    |    |-- struct_key: string (nullable = true)
|    |    |-- name: string (nullable = true)
{
role1_output: [
{
"struct_key": "test1",
"name": "person1"
}, 
{
"struct_key": "test1",
"name": "person2"
},
{
"struct_key": "test2",
"name": "person4"
},
{
"struct_key": "test2",
"name": "person5"
}
]
}
{
role2_output: [
{
"struct_key": "test1",
"name": "person3"
}, 
{
"struct_key": "test2",
"name": "person6"
}
]
}

I have tried the struct to map type conversion from that StackOverflow post answer but cannot figure out how to combine the extracted struct_key with another array of struct field values and create a new struct as I start transforming the array element, I lose the struct_key field value. Any advice?

答案1

得分: 0

见下文:

from pyspark.sql.functions import explode
from pyspark.sql.functions import array
from pyspark.sql.functions import struct
from pyspark.sql.functions import collect_list

输入 = {
    "a": [{
        "two": [{
            "name": "person1",
            "role": "role1"
        },
        {
            "name": "person2",
            "role": "role1"
        },
        {
            "name": "person3",
            "role": "role2"
        }
        ],
        "struct_key": "test1"
    },
    {
        "two": [{
            "name": "person4",
            "role": "role1"
        },
        {
            "name": "person5",
            "role": "role1"
        },
        {
            "name": "person6",
            "role": "role2"
        }
        ],
        "struct_key": "test2"
    }
    ]
}

df = spark.read.json(sc.parallelize([input])).selectExpr('inline(a)').select('struct_key', explode('two')).groupBy('col.role').agg(collect_list(struct('col.name','struct_key')))
df.show(truncate=False)
df.printSchema()
英文:

See below:

from pyspark.sql.functions import explode
from pyspark.sql.functions import array
from pyspark.sql.functions import struct
from pyspark.sql.functions import collect_list
input ={
"a": [{
"two": [{
"name": "person1",
"role": "role1"
},
{
"name": "person2",
"role": "role1"
},
{
"name": "person3",
"role": "role2"
}
],
"struct_key": "test1"
},
{
"two": [{
"name": "person4",
"role": "role1"
},
{
"name": "person5",
"role": "role1"
},
{
"name": "person6",
"role": "role2"
}
],
"struct_key": "test2"
}
]
}
df = spark.read.json(sc.parallelize([input])).selectExpr('inline(a)').select('struct_key', explode('two')).groupBy('col.role').agg(collect_list(struct('col.name','struct_key')))
df.show(truncate=False)
df.printSchema()

Gives you:

>>> df.show(truncate=False)
+-----+------------------------------------------------------------------------+
|role |collect_list(struct(col.name, struct_key))                              |
+-----+------------------------------------------------------------------------+
|role2|[{person3, test1}, {person6, test2}]                                    |
|role1|[{person1, test1}, {person2, test1}, {person4, test2}, {person5, test2}]|
+-----+------------------------------------------------------------------------+
>>> df.printSchema()
root
|-- role: string (nullable = true)
|-- collect_list(struct(col.name, struct_key)): array (nullable = false)
|    |-- element: struct (containsNull = false)
|    |    |-- name: string (nullable = true)
|    |    |-- struct_key: string (nullable = true)

huangapple
  • 本文由 发表于 2023年5月13日 11:49:33
  • 转载请务必保留本文链接:https://go.coder-hub.com/76240980.html
匿名

发表评论

匿名网友

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen:

确定