You can use selectExpr to make it a little cleaner:
df.selectExpr("cast(year as int) as year", "upper(make) as make","model", "comment", "blank")↧
Answer by dnlbrky for How can I change column types in Spark SQL's DataFrame?
↧