Data skewness in Spark leads to performance issues due to uneven partition distribution.
Dynamic partitioning in Hive allows for on-the-fly partition creation during data insertion.
Coalesce reduces partitions without shuffle; repartition changes partition count with shuffle.