你可以用 create_map 要创建Map列，请执行以下操作：

from pyspark.sql import functions as F

df1 = df.select(*[F.create_map(F.lit("v"), F.col(c)).alias(c) for c in df.columns])

df1.show()
+--------+------------+---------+
|      id|        Name|      Age|
+--------+------------+---------+
|[v -> 1]| [v -> John]|[v -> 31]|
|[v -> 2]|  [v -> Sam]|[v -> 34]|
|[v -> 3]|[v -> Chris]|[v -> 28]|
+--------+------------+---------+

使用 to_json 要从Map中获取json字符串：

df1.select(*[F.to_json(c).alias(c) for c in df1.columns]).show()

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

# |     id|         Name|     Age|

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

# |{"v":1}| {"v":"John"}|{"v":31}|

# |{"v":2}|  {"v":"Sam"}|{"v":34}|

# |{"v":3}|{"v":"Chris"}|{"v":28}|

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

您可以将每一列转换为一个具有名为 value ，并转换为json:

import pyspark.sql.functions as F

df2 = df.select([F.to_json(F.struct(F.col(c).alias('value'))).alias(c) for c in df.columns])

df2.show()
+-------------+-----------------+--------------+
|           id|             Name|           Age|
+-------------+-----------------+--------------+
|{"value":"1"}| {"value":"John"}|{"value":"31"}|
|{"value":"2"}|  {"value":"Sam"}|{"value":"34"}|
|{"value":"3"}|{"value":"Chris"}|{"value":"28"}|
+-------------+-----------------+--------------+