pyspark 笔记 cast 转换列的类型

1 不借助 pyspark.sql.types

from pyspark.sql.functions import col

data = [("Alice", "28"), ("Bob", "22"), ("Charlie", "30")]
columns = ["name", "age_str"]
df = spark.createDataFrame(data, columns)
df
#DataFrame[name: string, age_str: string]
#创建一个pyspark的DataFrame



#########################以上是源数据，以下是cast之后的结果############################
df.withColumn('cast_col',col('age_str').cast('int'))
#DataFrame[name: string, age_str: string, cast_col: int]

df.withColumn('cast_col',col('age_str').cast('float'))
#DataFrame[name: string, age_str: string, cast_col: float]

2 借助pysparks.sql.types

from pyspark.sql.types import *
df.withColumn('cast_col',col('age_str').cast(BooleanType()))
#DataFrame[name: string, age_str: string, cast_col: boolean]

|---------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------|
| BooleanType | |
| ByteType | 字节数据类型，占用一个字节的存储空间 |
| DateType | datetime.date 的数据类型 |
| DecimalType | 这个类型有两个可选参数，分别是 * precision------最大位数 * scale------小数点右侧位数 |
| DoubleType | |
| FloatType | |
| IntegerType | |
| LongType | |
| NullType | |
| ShortType | |
| StringType | |
| TimestampType | datetime.datetime 类型 |
| DayTimeIntervalType | datetime.timedelta类型 |