07 从RDD创建DataFrame

1.pandas df 与 spark df的相互转换

df_s=spark.createDataFrame(df_p)

df_p=df_s.toPandas()

 

2. Spark与Pandas中DataFrame对比

http://www.lining0806.com/spark%E4%B8%8Epandas%E4%B8%ADdataframe%E5%AF%B9%E6%AF%94/

 

3.1 利用反射机制推断RDD模式

• sc创建RDD
• 转换成Row元素，列名=值
• spark.createDataFrame生成df
• df.show(), df.printSchema()

 

from pyspark.sql import Row
people=spark.sparkContext.textFile('file:///usr/local/spark/mycode/rdd/chapter4-data01.txt').map(lambda line:line.split(',')).map(lambda p:Row(name=p[0],course=p[1],score=int(p[2])))
df=spark.createDataFrame(people)
df.first()
df.printSchema()

 

 

3.2 使用编程方式定义RDD模式

• 生成“表头”
  • fields = [StructField(field_name, StringType(), True) ,...]
  • schema = StructType(fields)
• 生成“表中的记录”
  • 创建RDD
  • 转换成Row元素，列名=值
• 把“表头”和“表中的记录”拼装在一起
  • = spark.createDataFrame(RDD, schema)

 

from pyspark.sql.types import *
from pyspark.sql import Row
schemaString='name course scroe'
fields=[StructField(field_name,StringType(),True) for field_name in schemaString.split(' ')]
schema=StructType(fields)
lines=spark.sparkContext.textFile('file:///usr/local/spark/mycode/rdd/chapter4-data01.txt')
parts=lines.map(lambda x:x.split(','))
people=parts.map(lambda p:Row(p[0],p[1],p[2].strip()))
schemaPeople=spark.createDataFrame(people,schema)
schemaPeople.show()

 

 

4. DataFrame保存为文件

schemaPeople.write.json("file:///usr/local/spark/schema_out")