使用 row_number 窗口函数 monotonically_increasing_id() ```
from pyspark.sql import *
from pyspark.sql.functions import *
w = Window.orderBy(monotonically_increasing_id())
df.withColumn("new_col",concat(split(col("my_string")," ")[0], lpad(row_number().over(w),2,"0"))).show()

+---------+-------+

|my_string|new_col|

+---------+-------+

|2020 test| 202001|

|2020 prod| 202002|

| 2020 dev| 202003|

+---------+-------+

``` UPDATE: 使用 when+otherwise 声明。

df.withColumn("dyn_col",when(lower(split(col("my_string")," ")[1]) =="prod","kk").\
when(lower(split(col("my_string")," ")[1]) =="dev","ff").\
when(lower(split(col("my_string")," ")[1]) =="test","01").\
otherwise("null")).\
withColumn("new_col",concat(split(col("my_string")," ")[0], col("dyn_col"))).\
drop("dyn_col").\
show()

# +---------+-------+

# |my_string|new_col|

# +---------+-------+

# |2020 test| 202001|

# |2020 prod| 2020kk|

# | 2020 dev| 2020ff|

# +---------+-------+

``` `In Scala:` ```
df.withColumn("dyn_col",when(lower(split(col("my_string")," ")(1)) ==="prod","kk").
when(lower(split(col("my_string")," ")(1)) ==="dev","ff").
when(lower(split(col("my_string")," ")(1)) ==="test","01").
otherwise("null")).
withColumn("new_col",concat(split(col("my_string")," ")(0), col("dyn_col"))).
drop("dyn_col").
show()

//+---------+-------+
//|my_string|new_col|
//+---------+-------+
//|2020 test| 202001|
//|2020 prod| 2020kk|
//| 2020 dev| 2020ff|
//+---------+-------+