How to pick fractional seconds in date pattern of spark sql?

I'm working in spark SQL and using this function to_date to convert the timestamp string in date format. But I'm unable to pick fractional seconds even though the documentation suggests so. Here it suggests that we can 'S' continuously, upto 6 singificant places, for fraction patterns.

For instance:

to_date("2017-09-09T21:01:12Z","yyyy-MM-dd'T'HH:mm:ss'Z'") -> working and returns date
to_date("2017-09-09T21:01:12.234433Z","yyyy-MM-dd'T'HH:mm:ss.SSSSSS'Z'") -> result in null

Can anybody suggests the possible reason for this behavior?

Edit (19-Apr-2023) : I believe this might be due to the version of spark that I'm using. Its not an offcial spark version and is probably a distribution of cloudera(I will confirm this) because the exact same function is properly working in official spark version 3.0.0

Two things should be changed (docs):

1. the format string is case-sensitive and
2. the character to use for the zone-offset with offset zero printed as Z is X

So the two format strings should be yyyy-MM-dd'T'HH:mm:ssX and yyyy-MM-dd'T'HH:mm:ss.SSSSSSX, but using yyyy-MM-dd'T'HH:mm:ss.SSSSSS'Z' also works of the zone-offset is always zero.