Try to change locale:

spark.read.format("csv").option("delimiter", ";").option("locale", "de")

By default it is en-US, see more details https://spark.apache.org/docs/3.5.4/sql-data-sources-csv.html#data-source-option