Spark SQL:将聚合函数应用于列列表 [英] Spark SQL: apply aggregate functions to a list of columns
问题描述
在执行 groupBy 时，有没有办法将聚合函数应用于数据帧的所有(或一系列)列?换句话说，有没有办法避免对每一列都这样做:
Is there a way to apply an aggregate function to all (or a list of) columns of a dataframe, when doing a groupBy? In other words, is there a way to avoid doing this for every column:
df.groupBy("col1")
.agg(sum("col2").alias("col2"), sum("col3").alias("col3"), ...)
推荐答案
有多种方法可以将聚合函数应用于多列.
There are multiple ways of applying aggregate functions to multiple columns.
GroupedData 类为最常用的函数提供了许多方法，包括count、max、min、mean和sum，可以直接使用如下:
GroupedData class provides a number of methods for the most common functions, including count, max, min, mean and sum, which can be used directly as follows:
Python:
Python:
df = sqlContext.createDataFrame(
[(1.0, 0.3, 1.0), (1.0, 0.5, 0.0), (-1.0, 0.6, 0.5), (-1.0, 5.6, 0.2)],
("col1", "col2", "col3"))
df.groupBy("col1").sum()
## +----+---------+-----------------+---------+
## |col1|sum(col1)| sum(col2)|sum(col3)|
## +----+---------+-----------------+---------+
## | 1.0| 2.0| 0.8| 1.0|
## |-1.0| -2.0|6.199999999999999| 0.7|
## +----+---------+-----------------+---------+
斯卡拉
Scala
val df = sc.parallelize(Seq(
(1.0, 0.3, 1.0), (1.0, 0.5, 0.0),
(-1.0, 0.6, 0.5), (-1.0, 5.6, 0.2))
).toDF("col1", "col2", "col3")
df.groupBy($"col1").min().show
// +----+---------+---------+---------+
// |col1|min(col1)|min(col2)|min(col3)|
// +----+---------+---------+---------+
// | 1.0| 1.0| 0.3| 0.0|
// |-1.0| -1.0| 0.6| 0.2|
// +----+---------+---------+---------+
可选地，您可以传递应聚合的列列表
Optionally you can pass a list of columns which should be aggregated
df.groupBy("col1").sum("col2", "col3")
您还可以将带有列 a 的键和函数作为值传递给字典/映射:
You can also pass dictionary / map with columns a the keys and functions as the values:
Python
Python
exprs = {x: "sum" for x in df.columns}
df.groupBy("col1").agg(exprs).show()
## +----+---------+
## |col1|avg(col3)|
## +----+---------+
## | 1.0| 0.5|
## |-1.0| 0.35|
## +----+---------+
斯卡拉
Scala
val exprs = df.columns.map((_ -> "mean")).toMap
df.groupBy($"col1").agg(exprs).show()
// +----+---------+------------------+---------+
// |col1|avg(col1)| avg(col2)|avg(col3)|
// +----+---------+------------------+---------+
// | 1.0| 1.0| 0.4| 0.5|
// |-1.0| -1.0|3.0999999999999996| 0.35|
// +----+---------+------------------+---------+
终于可以使用可变参数了:
Finally you can use varargs:
Python
Python
from pyspark.sql.functions import min
exprs = [min(x) for x in df.columns]
df.groupBy("col1").agg(*exprs).show()
斯卡拉
Scala
import org.apache.spark.sql.functions.sum
val exprs = df.columns.map(sum(_))
df.groupBy($"col1").agg(exprs.head, exprs.tail: _*)
还有其他一些方法可以达到类似的效果，但大多数情况下这些方法应该绰绰有余.
There are some other way to achieve a similar effect but these should more than enough most of the time.
另见:
这篇关于Spark SQL:将聚合函数应用于列列表的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
