Release notes for 0.193 #9802

nezihyigitbasi · 2018-01-23T23:50:58Z

David Phillips

01-17 12:32:23 5e2e9c4 Remove redundant throws clause from main code
01-17 17:02:50 59d7e03 Make default JDBC read mappings reusable
01-19 08:54:32 3a3fcbe Delete bytecode library
01-19 08:54:32 c26fa88 Extract bytecode library to separate project

Grzegorz Kokosiński

Haozhun Jin

01-17 17:44:06 a024eba Fix query failure when lambda is in UNNEST/VALUES
01-17 17:51:27 1536ce5 Make CLI and JDBC driver respect cookies
01-19 16:44:43 9107bd9 Improve readability in AbstractResourceConfigurationManager
01-22 15:44:48 e799922 Fix CLI to abort query when pager exits

James Sun

01-12 17:47:22 fa17537 Use page builder for array_remove
01-12 18:43:19 360b845 Support int.class in StateCompiler
01-12 18:46:01 4ca66c4 Introduce ObjectBlockPositionState and SliceBlockPositionState
01-12 18:46:02 6b147f7 Optimize arbitrary() with Object/SliceBlockPositionState
01-17 15:51:46 885d231 Fix stack overflow when toFailure is invoked with a cyclic throwable

Martin Traverso

01-18 10:37:32 b08ec02 Implement normal CDF to function
01-18 10:56:23 0fb527a Add SetDigest type and related functions
01-22 09:42:26 1816553 Remove deprecated optimizers

Nezih Yigitbasi

Piotr Findeisen

Wenlei Xie

01-12 21:34:35 bedf144 Make getRegion always return a new Block for structural block
01-22 18:07:12 57e8a1c Enforce instance factory MethodHandle not to have parameter
01-22 18:07:12 86ce0fc Use PageBuilder as state class for zip_with
01-22 18:07:12 d1a42b7 Add map_zip_with function

Ying Su

01-12 19:53:21 3faaad9 SetBuildOperator to support yielding when memory limit is exceeded

The text was updated successfully, but these errors were encountered:

wenleix · 2018-01-23T23:55:43Z

General

Add map_zip_with lambda function.

findepi · 2018-01-24T09:56:09Z

General

Fix Tried to free more revocable memory than is reserved error for queries that use spill and are moved to reserved memory pool.
Remove information_schema.__internal_partitions__ table. The table was useful for Hive connector only and became obsolete when SHOW PARTITIONS got improved.

Hive

Enable SHOW PARTITIONS FROM <table> WHERE <condition> to work for tables having more than hive.max-partitions-per-scan partitions as long as given condition reduces number of considered partitions below that limit.
Fix a bug causing ConcurrentModificationException when dropping any other column than the last one.
Fix a bug causing new partition to have incorrect on-disk data format in case of a partitioned table with non-default text format

findepi · 2018-01-24T09:58:17Z

dd60cdc Make read transformations extensible in base-jdbc

This is not user-facing nor a SPI change. @electrum should I write about this?

electrum · 2018-01-24T17:13:16Z

@findepi nope, no need to document this

JDBC Driver Changes
-------------------

* Use SSL by default for port 443.

nezihyigitbasi · 2018-01-24T21:26:03Z

General:

Add the peak of total non-revocable memory reservation (user + system memory) to query stats.

CLI Changes

Fix rendering maps of Bing tiles in the CLI.

kokosing · 2018-01-25T09:32:28Z

General:

Fix bug in iterative optimizer which could make planning to be endless. That was affecting queries with several predicates which were not eligible to be pushed down to connector.
Optimize query of partitioned table with a predicate that does not match to any partition to an empty relation.

comment to above:
Above is provide by e6d963b. f2b7f88 might look like a bug fix, but without e6d963b there was no such situation that partitioned table was pruned.

Security changes:

When using multiple authentication types, authentication methods are applied in order as they were specified in http-server.authentication.type.

Blackhole:

Add support for CREATE SCHEMA.
Fix bug which allowed to CREATE TABLE in schema which does not exist.

MongoDB:

Fix INSERT INTO.

Memory:

Fix bug which allowed to CREATE TABLE in schema which does not exist.
Fix bug in ALTER TABLE ... RENAME TO ... so now it is proper schema where renaming a table.

mbasmanova · 2018-01-26T19:07:17Z

Geospatial Functions:

Improve the performance of :func:ST_Contains.

nezihyigitbasi · 2018-01-26T21:05:22Z

Hive Changes

Fix a bug in PrestoS3FileSystem that causes queries to fail if they read data from regions other than us-east-1. Previously, such queries would fail with "The authorization header is malformed; the region 'us-east-1' is wrong; expecting '<region_name>'" , where <region_name> is the S3 region hosting the bucket that is queried.

nezihyigitbasi · 2018-01-26T23:21:01Z

@kokosing

Regarding general changes:

Fix bug in iterative optimizer which could make planning to be endless. -> Just to confirm, was the planner going into an infinite loop?
Optimize query of partitioned table with a predicate that does not match to any partition to an empty relation. -> I didn't understand this, can you please give more details?

Regarding memory connector, Fix bug in ALTER TABLE ... RENAME TO ... so now it is proper schema where renaming a table., can you please elaborate? I didn't understand the part about the schema.

Regarding mongodb Fix INSERT INTO., can you please elaborate? This is too generic.

martint · 2018-01-26T23:39:05Z

* Add :func:`normal_cdf`.
* Add ``SetDigest`` type and related functions.

yingsu00 · 2018-01-26T23:47:01Z

General:

Fix a wrong result issue on planner for row_number() with negative or zero filter values.
Support yielding for SetBuildOperator when memory limit is exceeded.

haozhun · 2018-01-27T00:06:18Z

General
-------
* Fix query analysis failure when lambda is used in ``UNNEST`` or ``VALUES``.

CLI Changes
-----------
* Fix CLI to abort the query when pager exists.

kokosing · 2018-01-29T11:30:50Z

@nezihyigitbasi

Fix bug in iterative optimizer which could make planning to be endless. -> Just to confirm, was the planner going into an infinite loop?

Yes.

Optimize query of partitioned table with a predicate that does not match to any partition to an empty relation. -> I didn't understand this, can you please give more details?

I think that example will be the most informative. Please take a look.
On 192 on hive with table partitioned on order_status:

presto:tpch> explain select * from test_insert_partitioned_table_existing_partition where order_status = 'x';
                                                                                                 Query Plan
--------------------------------------------------------------------------------------------------------------------------------------------------------------
 - Output[order_key, comment, order_status] => [order_key:bigint, comment:varchar, order_status:varchar]
     - RemoteExchange[GATHER] => order_key:bigint, comment:varchar, order_status:varchar
         - TableScan[hive:tpch:test_insert_partitioned_table_existing_partition, originalConstraint = ("order_status" = CAST('x' AS varchar))] => [order_key:b
                 LAYOUT: tpch.test_insert_partitioned_table_existing_partition
                 :: NONE

Notice NONE which suggests that no partition is going be read.

While in current master:

presto:tpch> explain select * from test_insert_partitioned_table_existing_partition where order_status = 'x';
                                               Query Plan
---------------------------------------------------------------------------------------------------------
 - Output[order_key, comment, order_status] => [order_key:bigint, comment:varchar, order_status:varchar]
         Cost: {rows: 0, bytes: ?}
     - Values => [order_key:bigint, comment:varchar, order_status:varchar]
             Cost: {rows: 0, bytes: ?}

Regarding memory connector, Fix bug in ALTER TABLE ... RENAME TO ... so now it is proper schema where renaming a table., can you please elaborate? I didn't understand the part about the schema.

Please see: #9688, previously when:

alter table memory.default.t1 rename to memory.default2.t2;

Previously with above, table default.t1 was renamed wrongly to default.t2, now it is renamed to default2.t2.

Regarding mongodb Fix INSERT INTO., can you please elaborate? This is too generic.

I am sorry, INSERT INTO did not work for mongodb at all, as there was no insert related method in connector, see #9736.

nezihyigitbasi mentioned this issue Jan 29, 2018

Add 0.193 release notes #9845

Merged

nezihyigitbasi closed this as completed Jan 30, 2018

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Release notes for 0.193 #9802

Release notes for 0.193 #9802

nezihyigitbasi commented Jan 23, 2018 •

edited

wenleix commented Jan 23, 2018

findepi commented Jan 24, 2018

findepi commented Jan 24, 2018

electrum commented Jan 24, 2018

nezihyigitbasi commented Jan 24, 2018

kokosing commented Jan 25, 2018

mbasmanova commented Jan 26, 2018 •

edited

nezihyigitbasi commented Jan 26, 2018

nezihyigitbasi commented Jan 26, 2018

martint commented Jan 26, 2018 •

edited

yingsu00 commented Jan 26, 2018

haozhun commented Jan 27, 2018

kokosing commented Jan 29, 2018

Release notes for 0.193 #9802

Release notes for 0.193 #9802

Comments

nezihyigitbasi commented Jan 23, 2018 • edited

David Phillips

Grzegorz Kokosiński

Haozhun Jin

James Sun

Martin Traverso

Nezih Yigitbasi

Piotr Findeisen

Wenlei Xie

Ying Su

wenleix commented Jan 23, 2018

General

findepi commented Jan 24, 2018

General

Hive

findepi commented Jan 24, 2018

electrum commented Jan 24, 2018

nezihyigitbasi commented Jan 24, 2018

General:

CLI Changes

kokosing commented Jan 25, 2018

General:

Security changes:

Blackhole:

MongoDB:

Memory:

mbasmanova commented Jan 26, 2018 • edited

Geospatial Functions:

nezihyigitbasi commented Jan 26, 2018

Hive Changes

nezihyigitbasi commented Jan 26, 2018

martint commented Jan 26, 2018 • edited

yingsu00 commented Jan 26, 2018

General:

haozhun commented Jan 27, 2018

kokosing commented Jan 29, 2018

nezihyigitbasi commented Jan 23, 2018 •

edited

mbasmanova commented Jan 26, 2018 •

edited

martint commented Jan 26, 2018 •

edited