In this region, for each field contained in the row, a bit is reserved to indicate whether it contains a null value or not. This is very useful for filtering, because only the bit set has to be read from memory, omitting the actual values. The number of fields and the size of this region are reflected as individual properties in the org.apache.spark.sql.catalyst.expressions.UnsafeRow object. Therefore, this region can be of variable size as well, which is an implicit requirement since the number of fields varies as well.