你可以在 countDistinct ，一个取决于列中是否有空值的整数0或1：

# example dataframe

df.show()
+---+---+----+----+
|  a|  b|   c|   d|
+---+---+----+----+
|  1|  1|   1|null|
|  1|  2|null|null|
+---+---+----+----+

import pyspark.sql.functions as F

cnt = df.agg(*[
    (F.countDistinct(c) + F.max(F.col(c).isNull()).cast('int')).alias(c) 
    for c in df.columns
])

cnt.show()
+---+---+---+---+
|  a|  b|  c|  d|
+---+---+---+---+
|  1|  2|  2|  1|
+---+---+---+---+

df.drop(*[c for c in cnt.first().asDict() if cnt.first()[c] == 1]).show()
+---+----+
|  b|   c|
+---+----+
|  1|   1|
|  2|null|
+---+----+