11.180. Release 0.61

Add support for Table Value Constructors

Presto now supports the SQL table value constructor syntax to create inline tables.The VALUES clause can be used anywhere a SELECT statement is allowed.For example, as a top-level query:

  1. VALUES ('a', 1), ('b', 2);
  1. _col0 | _col1
  2. -------+-------
  3. a | 1
  4. b | 2
  5. (2 rows)

Alternatively, in the FROM clause:

  1. SELECT *
  2. FROM (
  3. VALUES
  4. ('a', 'ape'),
  5. ('b', 'bear')
  6. ) AS animal (letter, animal)
  7. JOIN (
  8. VALUES
  9. ('a', 'apple'),
  10. ('b', 'banana')
  11. ) AS fruit (letter, fruit)
  12. USING (letter);
  1. letter | animal | letter | fruit
  2. --------+--------+--------+---------
  3. a | ape | a | apple
  4. b | bear | b | banana
  5. (2 rows)

Cassandra

  • Add support for upper-case schema, table, and columns names.
  • Add support for DECIMAL type.

Amazon S3 support

  • Completely rewritten Hadoop FileSystem implementation for S3 using the Amazon AWS SDK,with major performance and reliability improvements.
  • Add support for writing data to S3.

Miscellaneous

  • General improvements to the JDBC driver, specifically with respect to metadata handling.
  • Fix division by zero errors in variance aggregation functions (VARIANCE, STDDEV, etc.).
  • Fix a bug when using DISTINCT aggregations in the HAVING clause.
  • Fix an out of memory issue when writing large tables.
  • Fix a bug when using ORDER BY rand() in a JOIN query.
  • Fix handling of timestamps in maps and lists in Hive connector.
  • Add instrumentation for Hive metastore and HDFS API calls to track failures and latency. These metrics are exposed via JMX.