PySpark – 按多列排序

在本文中，我们将了解如何通过Python对 PySpark DataFrames 中的多列进行排序。

创建用于演示的数据框：

Python3

# importing module
import pyspark
  
# importing sparksession from pyspark.sql module
from pyspark.sql import SparkSession
  
# creating sparksession and giving an app name
spark = SparkSession.builder.appName('sparkdf').getOrCreate()
  
# list  of employee data
data = [["1", "sravan", "company 1"],
        ["2", "ojaswi", "company 1"],
        ["3", "rohith", "company 2"],
        ["4", "sridevi", "company 1"],
        ["5", "bobby", "company 1"]]
  
# specify column names
columns = ['ID', 'NAME', 'Company']
  
# creating a dataframe from the lists of data
dataframe = spark.createDataFrame(data, columns)
  
dataframe.show()

输出：

orderby 表示我们将按多列按升序或降序对数据帧进行排序。我们可以使用以下方法来做到这一点。

方法一：使用 orderBy()

此函数将在对多列进行排序后返回数据框。它将首先根据给定的列名进行排序。

Syntax:

• Ascending order: dataframe.orderBy([‘column1′,’column2′,……,’column n’], ascending=True).show()
• Descending Order: dataframe.orderBy([‘column1′,’column2′,……,’column n’], ascending=False).show()

where:

• dataframe is the Pyspark Input dataframe
• ascending=True specifies to sort the dataframe in ascending order
• ascending=False specifies to sort the dataframe in descending order

示例 1 ：使用 orderBy() 按升序对 PySpark 数据帧进行排序。

Python3

# importing module
import pyspark
  
# importing sparksession from pyspark.sql module
from pyspark.sql import SparkSession
  
# creating sparksession and giving an app name
spark = SparkSession.builder.appName('sparkdf').getOrCreate()
  
# list  of employee data
data = [["1", "sravan", "company 1"],
        ["2", "ojaswi", "company 1"],
        ["3", "rohith", "company 2"],
        ["4", "sridevi", "company 1"],
        ["5", "bobby", "company 1"]]
  
# specify column names
columns = ['ID', 'NAME', 'Company']
  
# creating a dataframe from the lists of data
dataframe = spark.createDataFrame(data, columns)
  
# orderBy dataframe in asc order
dataframe.orderBy(['Name', 'ID', 'Company'],
                  ascending=True).show()

输出：

示例 2：使用 orderBy() 对 PySpark 数据帧进行降序排序。

Python3

# importing module
import pyspark
  
# importing sparksession from pyspark.sql module
from pyspark.sql import SparkSession
  
# creating sparksession and giving an app name
spark = SparkSession.builder.appName('sparkdf').getOrCreate()
  
# list  of employee data
data = [["1", "sravan", "company 1"],
        ["2", "ojaswi", "company 1"],
        ["3", "rohith", "company 2"],
        ["4", "sridevi", "company 1"],
        ["5", "bobby", "company 1"]]
  
# specify column names
columns = ['ID', 'NAME', 'Company']
  
# creating a dataframe from the lists of data
dataframe = spark.createDataFrame(data, columns)
  
# orderBy dataframe in desc order
dataframe.orderBy(['Name', 'ID', 'Company'], 
                  ascending=False).show()

输出：

方法 2：使用 sort()

此函数将在对多列进行排序后返回数据框。它将首先根据给定的列名进行排序。

Syntax:

• Ascending order: dataframe.sort([‘column1′,’column2′,……,’column n’], ascending=True).show()
• Descending Order: dataframe.sort([‘column1′,’column2′,……,’column n’], ascending=False).show()

where,

1. dataframe is the Pyspark Input dataframe
2. ascending=True specifies to sort the dataframe in ascending order
3. ascending=False specifies to sort the dataframe in descending order

示例 1：按升序对 PySpark 数据帧进行排序

Python3

# importing module
import pyspark
  
# importing sparksession from pyspark.sql module
from pyspark.sql import SparkSession
  
# creating sparksession and giving an app name
spark = SparkSession.builder.appName('sparkdf').getOrCreate()
  
# list  of employee data
data = [["1", "sravan", "company 1"],
        ["2", "ojaswi", "company 1"],
        ["3", "rohith", "company 2"],
        ["4", "sridevi", "company 1"],
        ["5", "bobby", "company 1"]]
  
# specify column names
columns = ['ID', 'NAME', 'Company']
  
# creating a dataframe from the lists of data
dataframe = spark.createDataFrame(data, columns)
  
# orderBy dataframe in asc order
dataframe.sort(['Name', 'ID', 'Company'],
               ascending=True).show()

输出：

示例 2：按降序对 PySpark 数据帧进行排序

Python3

# importing module
import pyspark
  
# importing sparksession from pyspark.sql module
from pyspark.sql import SparkSession
  
# creating sparksession and giving an app name
spark = SparkSession.builder.appName('sparkdf').getOrCreate()
  
# list  of employee data
data = [["1", "sravan", "company 1"],
        ["2", "ojaswi", "company 1"],
        ["3", "rohith", "company 2"],
        ["4", "sridevi", "company 1"],
        ["5", "bobby", "company 1"]]
  
# specify column names
columns = ['ID', 'NAME', 'Company']
  
# creating a dataframe from the lists of data
dataframe = spark.createDataFrame(data, columns)
  
# orderBy dataframe in desc order
dataframe.sort(['Name', 'ID', 'Company'],
               ascending=False).show()

输出：