尝试从 pyspark 中的数据框中删除嵌套列不起作用。
这是我的代码中的 sinppet:
`from pyspark.sql import functions as F from pyspark.sql.types import IntegerType , BooleanType from pyspark.sql.functions import udf
#simple filter function
@F.udf(returnType=BooleanType())
def my_filter(x):
if (x != df_new.a.b) :
return True
else :
return False
#df.filter(my_filter('id', 'category')).show()
def drop_col(df, struct_nm, delete_struct_child_col_nm):
#fields_to_keep = filter(lambda x: x != delete_struct_child_col_nm, df.select("{}.*".format(struct_nm)).columns)
fields_to_keep = filter(lambda x: my_filter(x) , df.select("{}.*".format(struct_nm)).columns)
fields_to_keep = list(map(lambda x: "{}.{}".format(struct_nm, x), fields_to_keep))
return df.withColumn(struct_nm, struct(fields_to_keep))
drop_col(df_new,“a”,df_new.ab)`
我使用 UDF 是因为尝试以下行不起作用。因为这不是!= 符号不起作用,也不能使用波浪号
#fields_to_keep = filter(lambda x: x != delete_struct_child_col_nm, df.select("{}.*".format
编辑:有人在评论中询问架构,所以我提供它
root
|-- a: struct (nullable = false)
| |-- rawEntity: string (nullable = true)
| |-- entityType: string (nullable = true)
| | |-- element: struct (containsNull = true)
| | | |-- StoreId: string (nullable = true)
| | | |-- AppId: string (nullable = true)
| |-- Timestamps: array (nullable = true)
| | |-- element: long (containsNull = true)
| |-- User: string (nullable = true)
| |-- b: array (nullable = true) //trying to drop this
| | |-- element: string (containsNull = true)
| |-- KeywordsFull: array (nullable = true)
| | |-- element: struct (containsNull = true)
| | | |-- Keywords: string (nullable = true)
| | |-- element: string (containsNull = true)