Some columns are not written to Redis

[iulianm82]

Hi,

I am new to redis and spark-redis.
I have a problem that I struggle to understand and solve, so I would appreciate a lot any help:

I am trying to write to redis a dataset with 99 rows and then read it. Both the writing and reading are successful, but the problem is that 3 out of 14 columns in the dataset are simply not written to redis. What these 3 columns have in common is that they contain a big number of nulls: one column contains only nulls, one column contains 93 and the third one contains 43 nulls.
The weird thing is that there are other 3 columns containing nulls but they are written to redis every time: one column with 51 nulls, one column with 31 nulls and one with 3 nulls.

This is the code for writing, pretty standard:
dataset.write()
.format("org.apache.spark.sql.redis")
.option("table", uniqueName)
.mode(SaveMode.Overwrite)
.save();
How spark-redis handles nulls and what I could do to have all columns be written to redis?

Thanks in advance,
Iulian