12.1. Release 334 (29 May 2020)
General Changes
- Fix incorrect query results for certain queries involving comparisons of
real
and double
types
when values include negative zero. (#3745)
- Fix failure when querying an empty table with late materialization enabled. (#3577)
- Fix failure when the inputs to
UNNEST
are repeated. (#3587)
- Fix failure when an aggregation is used in the arguments to
format()
. (#3829)
- Fix
localtime()
and current_time()
for session zones with DST or with historical offset changes
in legacy (default) timestamp semantics. (#3846, #3850)
- Fix dynamic filter failures in complex spatial join queries. (#3694)
- Improve performance of queries involving
row_number()
. (#3614)
- Improve performance of queries containing
LIKE
predicate. (#3618)
- Improve query performance when dynamic filtering is enabled. (#3632)
- Improve performance for queries that read fields from nested structures. (#2672)
- Add variant of
random()
function that produces a number in the provided range. (#1848)
- Show distributed plan by default in EXPLAIN. (#3724)
- Add Oracle Connector. (#1959)
- Add Pinot Connector. (#2028)
- Add Prometheus Connector. (#2321)
- Add support for standards compliant (RFC 7239) HTTP forwarded headers. Processing of HTTP forwarded headers is now controlled by the
http-server.process-forwarded
configuration property, and the old http-server.authentication.allow-forwarded-https
and
dispatcher.forwarded-header
configuration properties are no longer supported. (#3714)
- Add pluggable Certificate Authenticator. (#3804)
JDBC Driver
- Implement
toString()
for java.sql.Array
results. (#3803)
CLI Changes
- Improve rendering of elapsed time for short queries. (#3311)
Web UI Changes
- Add
fixed
, certificate
, JWT
, and Kerberos
to UI authentication. (#3433)
- Show join distribution type in Live Plan. (#1323)
JDBC Driver Changes
- Improve performance of
DatabaseMetaData.getColumns()
when the
parameters contain unescaped %
or _
. (#1620)
Elasticsearch Connector Changes
- Fix failure when executing
SHOW CREATE TABLE
. (#3718)
- Improve performance for
count(*)
queries. (#3512)
- Add support for raw Elasticsearch queries. (#3735)
Hive Connector Changes
- Fix matching bucket filenames without leading zeros. (#3702)
- Fix creation of external tables using
CREATE TABLE AS
. Previously, the
tables were created as managed and with the default location. (#3755)
- Fix incorrect table statistics for newly created external tables. (#3819)
- Prevent Presto from starting when cache fails to initialize. (#3749)
- Fix race condition that could cause caching to be permanently disabled. (#3729, #3810)
- Fix malformed reads when asynchronous read mode for caching is enabled. (#3772)
- Fix eviction of cached data while still under size eviction threshold. (#3772)
- Improve performance when creating unpartitioned external tables over large data sets. (#3624)
- Leverage Parquet file statistics when reading decimal columns. (#3581)
- Change type of
$file_modified_time
hidden column from bigint
to timestamp with timezone type
. (#3611)
- Add caching support for HDFS and Azure file systems. (#3772)
- Fix S3 connection pool depletion when asynchronous read mode for caching is enabled. (#3772)
- Disable caching on coordinator by default. (#3820)
- Use asynchronous read mode for caching by default. (#3799)
- Cache delegation token for Hive thrift metastore. This can be configured with
the
hive.metastore.thrift.delegation-token.cache-ttl
and hive.metastore.thrift.delegation-token.cache-maximum-size
configuration properties. (#3771)
MongoDB Connector Changes
- Support case insensitive database and collection names. This can be enabled with the
mongodb.case-insensitive-name-matching
configuration property. (#3453)
SPI Changes
- Allow a
SystemAccessControl
to provide an EventListener
. (#3629).