记录备忘:
转自: https://www.jianshu.com/p/177cbcb1cb6f
数据拉取
加载包:
from __future__ import print_function
import pandas as pd
from pyspark.sql import HiveContext
from pyspark import SparkContext,SparkConf
from sqlalchemy import create_engine
import datetime
import pyspark.sql.functions as F
conf = SparkConf().setAppName("abc")
sc = SparkContext(conf=conf)
hiveCtx = HiveContext(sc)
# 创建dataframe
d = [{'name': 'Alice', 'age': 1},{'name': 'Bob', 'age': 5}]
df = sqlContext.createDataFrame(d)
df.show()
sql = "" # 拉数SQL
df = hiveCtx.sql(sql)
数据探索
df.show() # 不加参数默认展示前20行
df.count()
df.printSchema()
df.columns
数据处理
df.select('age','name') # 带show才能看到结果
df.select(df.age.alias('age_value'),'name')
df.filter(df.name=='Alice')
函数和UDF
pyspark.sql.functions里有许多常用的函数,可以满足日常绝大多数的数据处理需求;当然也支持自己写的UDF,直接拿来用。
自带函数
根据官方文档,以下是部分函数说明:
'lit': 'Creates a :class:`Column` of literal value.',
'col': 'Returns a :class:`Column` based on the given column name.',
'column': 'Returns a :class:`Column` based on the given column name.',
'asc': 'Returns a sort expression based on the ascending order of the given column name.',
'desc': 'Returns a sort expression based on the descending order of the given column name.',
'upper': 'Converts a string expression to upper case.',
'lower': 'Converts a string expression to upper case.',
'sqrt': 'Computes the square root of the specified float value.',
'abs': 'Computes the absolutle value.',
'max': 'Aggregate function: returns the maximum value of the expression in a group.',
'min': 'Aggregate function: returns the minimum value of the expression in a group.',
'first': 'Aggregate function: returns the first value in a group.',
'last': 'Aggregate function: returns the last value in a group.',
'count': 'Aggregate function: returns the number of items in a group.',
'sum': 'Aggregate function: returns the sum of all values in the expression.',
'avg': 'Aggregate function: returns the average of the values in a group.',
'mean': 'Aggregate function: returns the average of the values in a group.',
'sumDistinct': 'Aggregate function: returns the sum of distinct values in the expression.',
---------------------------
df.select(F.max(df.age))
df.select(F.min(df.age))
df.select(F.avg(df.age)) # 也可以用mean,一样的效果
df.select(F.countDistinct(df.age)) # 去重后统计
df.select(F.count(df.age)) # 直接统计,经试验,这个函数会去掉缺失值会再统计
from pyspark.sql import Window
df.withColumn("row_number", F.row_number().over(Window.partitionBy("a","b","c","d").orderBy("time"))).show() # row_number()函数
数据写出
写入集群分区表
all_bike.rdd.map(lambda line: u','.join(map(lambda x:unicode(x),line))).saveAsTextFile('/user/hive/warehouse/bi.db/bikeid_without_3codes_a_d/dt={}'.format(t0_uf)) #转化为RDD写入HDFS路径
还有一种方法,是先把dataframe创建成一个临时表,再用hive sql的语句写入表的分区
bike_change_2days.registerTempTable('bike_change_2days')
sqlContext.sql("insert into bi.bike_changes_2days_a_d partition(dt='%s') select citycode,biketype,detain_bike_flag,bike_tag_onday,bike_tag_yesterday,bike_num from bike_change_2days"%(date))
写入集群非分区表
df_spark.write.mode("append").insertInto('bi.pesudo_bike_white_list') # 直接使用write.mode方法insert到指定的集群表
可以先将PySpark DataFrame转化成Pandas DataFrame,然后用pandas的to_sql方法插入数据库
写出本地
df.write.csv()
与Pandas DataFrame互相转换
如果你熟悉Pandas包,并且PySpark处理的中间数据量不是太大,那么可以直接转换成pandas DataFrame,然后转化成常规操作。
df.toPandas() # PySpark DataFrame转化成Pandas DataFrame
import pandas as pd
df_p = pd.DataFrame(dict(num=range(3),char=['a','b','c']))
df_s = sqlContext.createDataFrame(df_p) # pandas dataframe转化成PySpark DataFrame
type(df_s)
来源:oschina
链接:https://my.oschina.net/u/4306166/blog/3234158