如何在pyspark数据帧中一行替换regexp_?
我有一个pyspark数据帧列如何在pyspark数据帧中一行替换regexp_?,pyspark,pyspark-dataframes,Pyspark,Pyspark Dataframes,我有一个pyspark数据帧列 df.groupBy('Gender').count().show() (5) Spark Jobs +------+------+ |Gender| count| +------+------+ | F| 44015| | null| 42175| | M|104423| | | 1| +------+------+ 我正在做regexp\u替换 #df = df.fillna({'Gender':'missing'}) d
df.groupBy('Gender').count().show()
(5) Spark Jobs
+------+------+
|Gender| count|
+------+------+
| F| 44015|
| null| 42175|
| M|104423|
| | 1|
+------+------+
我正在做regexp\u替换
#df = df.fillna({'Gender':'missing'})
df = df.withColumn('Gender', regexp_replace('Gender', 'F','Female'))
df = df.withColumn('Gender', regexp_replace('Gender', 'M','Male'))
df = df.withColumn('Gender', regexp_replace('Gender', ' ','missing'))
不是为每行调用df,而是在一行中完成吗?如果您不想使用
regexp\u replace
三次,您可以在/when/other子句时使用
from pyspark.sql import functions as F
from pyspark.sql.functions import when
df.withColumn("Gender", F.when(F.col("Gender")=='F',F.lit("Female"))\
.when(F.col("Gender")=='M',F.lit("Male"))\
.otherwise(F.lit("missing"))).show()
+-------+------+
| Gender| count|
+-------+------+
| Female| 44015|
|missing| 42175|
| Male|104423|
|missing| 1|
+-------+------+
或者,您可以在一行中替换三个regexp\u
,如下所示:
from pyspark.sql.functions import regexp_replace
df.withColumn('Gender', regexp_replace(regexp_replace(regexp_replace('Gender', 'F','Female'),'M','Male'),' ','missing')).show()
+-------+------+
| Gender| count|
+-------+------+
| Female| 44015|
| null| 42175|
| Male|104423|
|missing| 1|
+-------+------+
我认为当/否则
应优于3个regexp\u replace
函数,因为您也需要使用fillna