pyspark读取pickle文件内容并存储到hive

在平常工作中，难免要和大数据打交道，而有时需要读取本地文件然后存储到Hive中，本文接下来将具体讲解。

过程：

使用pickle模块读取.plk文件；
将读取到的内容转为RDD；
将RDD转为DataFrame之后存储到Hive仓库中；

1、使用pickle保存和读取pickle文件

import pickle
data = ""
path = "xxx.plj"
#保存为pickle
pickle.dump(data,open(path,'wb'))
#读取pickle
data2 = pickle.load(open(path,'rb'))

使用python3读取python2保存的pickle文件时，会报错：

UnicodeDecodeError: 'ascii' codec can't decode byte 0xa0 in position 11: ordinal not in range(128)

解决方法：

data2 = pickle.load(open(path,'rb',encoding='latin1'))

使用python2读取python3保存的pickle文件时，会报错：

unsupported pickle protocol:3

解决方法：

import pickle
path = "xxx.plk"
path2 = 'xxx2.plk'
data = pickle.load(open(path,'rb'))
#保存为python2的pickle
pickle.dump(data,open(path2,'wb'),protocol=2)
#读取pickle
data2 = pickle.load(open(path2,'rb'))

2、读取pickle的内容并转为RDD

from pyspark.sql import SparkSession
from pyspark.sql import Row
import pickle


spark = SparkSession 
    .builder 
    .appName("Python Spark SQL basic example") 
    .config("spark.some.config.option", "some-value") 
    .getOrCreate()
with open(picle_path,"rb") as fp:
    data = pickle.load(fp)
    #这里可根据data的类型进行相应的操作

#假设data是一个一维数组：[1,2,3,4,5]，读取数据并转为rdd
pickleRdd = spark.parallelize(data)

3、将rdd转为dataframe并存入到Hive中

#定义列名
column = Row('col')
#转为dataframe
pickleDf =pickleRdd.map(lambda x:column(x))
#存储到Hive中，会新建数据库：hive_database，新建表：hive_table，以覆盖的形式添加，partitionBy用于指定分区字段
pickleDf..write.saveAsTable("hive_database.hvie_table", mode='overwrite', partitionBy=‘’)

补充存入到Hive中的知识：

（1）通过sql的方式

data = [
    (1,"3","145"),
    (1,"4","146"),
    (1,"5","25"),
    (1,"6","26"),
    (2,"32","32"),
    (2,"8","134"),
    (2,"8","134"),
    (2,"9","137")
]
df = spark.createDataFrame(data, ['id', "test_id", 'camera_id'])
 
# method one，default是默认数据库的名字，write_test 是要写到default中数据表的名字
df.registerTempTable('test_hive')
sqlContext.sql("create table default.write_test select * from test_hive")

或者：

# df 转为临时表/临时视图
df.createOrReplaceTempView("df_tmp_view")
# spark.sql 插入hive
spark.sql(""insert overwrite table 
                    XXXXX  # 表名
                   partition(分区名称=分区值)   # 多个分区按照逗号分开
                   select 
                   XXXXX  # 字段名称，跟hive字段顺序对应，不包含分区字段
                   from df_tmp_view""")

（2）以saveAsTable的形式

# "overwrite"是重写表的模式，如果表存在，就覆盖掉原始数据，如果不存在就重新生成一张表
#  mode("append")是在原有表的基础上进行添加数据
df.write.format("hive").mode("overwrite").saveAsTable('default.write_test')

以下是通过rdd创建dataframe的几种方法：

（1）通过键值对

d = [{'name': 'Alice', 'age': 1}]
output = spark.createDataFrame(d).collect()
print(output)

# [Row(age=1, name='Alice')]

（2）通过rdd

a = [('Alice', 1)]
rdd = sc.parallelize(a)
output = spark.createDataFrame(rdd).collect()
print(output)
output = spark.createDataFrame(rdd, ["name", "age"]).collect()
print(output)

# [Row(_1='Alice', _2=1)]
# [Row(name='Alice', age=1)]

（3）通过rdd和Row

from pyspark.sql import Row


a = [('Alice', 1)]
rdd = sc.parallelize(a)
Person = Row("name", "age")
person = rdd.map(lambda r: Person(*r))
output = spark.createDataFrame(person).collect()
print(output)

# [Row(name='Alice', age=1)]

（4）通过rdd和StrutType

from pyspark.sql.types import *

a = [('Alice', 1)]
rdd = sc.parallelize(a)
schema = StructType(
    [
        StructField("name", StringType(), True),
        StructField("age", IntegerType(), True)
    ]
)
output = spark.createDataFrame(rdd, schema).collect()
print(output)

# [Row(name='Alice', age=1)]

（5）基于pandas dataframe创建

df = spark.createDataFrame(rdd, ['name', 'age'])
print(df)  # DataFrame[name: string, age: bigint]

print(type(df.toPandas()))  # <class 'pandas.core.frame.DataFrame'>

# 传入pandas DataFrame
output = spark.createDataFrame(df.toPandas()).collect()
print(output)

# [Row(name='Alice', age=1)]

参考：

https://blog.csdn.net/sinat_28224453/article/details/84977693

https://blog.csdn.net/weixin_39198406/article/details/104916715

https://blog.csdn.net/u011412768/article/details/93426353