通过以下代码：
SELECT
farm_products.id as ID,
farm_products.product as Product1,
shop_products.product as Product2
FROM
farm_products,
shop_products,
shop_farm
WHERE
farm_products.id = shop_farm.farm_id and
shop_farm.farm_id = shop_products.id and
farm_products.product != shop_products.product;
我得到以下输出：
+-------+----------+---------+
|ID | Product1 | Product2|
+-------+----------+---------+
|06 | 'Apple' | 'Grape' |
+-------+----------+---------+
|06 | 'Orange' | 'Grape' |
+-------+----------+---------+
|06 | 'Pear' | 'Apple' |
+-------+----------+---------+
|07 | 'Apple' | 'Pear' |
+-------+----------+---------+
|08 | 'Kiwi' | 'Grape' |
+-------+----------+---------+
|08 | 'Grape' | 'Orange |
+-------+----------+---------+
我想要一个表，其中每个表仅出现第一行
出现ID。
换句话说，我想要一个看起来像这样的输出：
+-------+----------+---------+
|ID | Product1 | Product2|
+-------+----------+---------+
|06 | 'Apple' | 'Grape' |
+-------+----------+---------+
|07 | 'Apple' | 'Pear' |
+-------+----------+---------+
|08 | 'Kiwi' | 'Grape' |
+-------+----------+---------+
我尝试使用 DISTINCT 删除所有重复的ID，但这（显然）不起作用。 我想尝试避免嵌套查询，并使代码尽可能简单。
有人可以帮忙吗？
答案 0 :(得分：1)
使用export class CustomComponent implements , ControlValueAccessor {
onChange = (val: string) => { };
onTouched = () => { };
writeValue(val: string): void {
// value passed from parent throug ngModel will come under this funtion
}
registerOnChange(fn: (val: string) => void): void {
this.onChange = fn;
}
registerOnTouched(fn: () => void): void {
this.onTouched = fn;
}
ngOnInit() {
}
// If you want to emit value to parent use the onChange function
myEmitFunction(){
this.onChange("value u want to emit")
}
}
row_number()您可以通过以下方式使用查询
select * from
(select *, row_number() over( partition by id order by Product1 ) rn
from table_name
)t where rn=1
答案 1 :(得分：0)
这将起作用：
from pyspark.sql.window import Window
from pyspark.sql.functions import rowNumber
w = Window().orderBy()
df = df.withColumn("rowid", rowNumber().over(w))
forecasts = forecasts.withColumn("rowid", rowNumber().over(w))
mergedDF = df.join(forecasts, "rowid").drop("rowid")
mergedDF.show()