

By

CREATE TABLE

PARTITIONED BY

[PARTITIONED BY (col3 data_type [COMMENT col_comment], ...)]

CLUSTERED/SORTED BY

[CLUSTERED BY (col1, ...) [SORTED BY (col1 [ASC|DESC], ...)] INTO num_buckets BUCKETS]

SKEWED BY

[SKEWED BY (col1, col2, ...) ON ((col_value, col_value, ...), ...)

Query

ORDER BY

CLUSTER BY

DISTRIBUTE BY(repartition)

# The default number of partitions to use when shuffling data for joins or aggregations.
spark.sql.shuffle.partitions=200

SORT BY

WINDOW

CLUSTER BY

PARTITION|DISTRIBUTE BY

ORDER|SORT BY

Writer

partitionBy

bucketBy

sortBy

Reference