Similar to SPARK-342 SPARK-340.

Use of Timestamp and Scala Arrays in where statements use the codec registry directly and throw because of missing codecs. Unlike the write side which uses RowToBsonDocumentConverter.

spark.read.format("mongodb").where("myArray == array('123')").collect()

spark.read.format("mongodb").where("createdDate = cast('2022-01-06T07:06:51.628' as timestamp)").collect()