drill

Clone Tools
  • last updated 29 mins ago
Constraints
Constraints: committers
 
Constraints: files
Constraints: dates
updates for drill 1.8 release

    • -31
    • +0
    /blog/_posts/2016-08-15-drill-1.8-released.md
    • -0
    • +30
    /blog/_posts/2016-08-30-drill-1.8-released.md
updates to docs for Drill 1.8 release

    • -4
    • +4
    /_docs/install/047-installing-drill-on-the-cluster.md
Update version to 1.9.0-SNAPSHOT.

    • -1
    • +1
    /contrib/data/tpch-sample-data/pom.xml
  1. … 12 more files in changeset.
update to 1.8 release notes

Use DBDocumentReaderBase instead of DBDocumentReader.

MD-290: Fix build failure due to upstream changes.

+ Includes MD-292

DRILL-4862: Fix binary_string to use an injected buffer as out buffer

+ Previously, binary_string used the input buffer as output buffer. So after calling binary_string, the original content was destroyed. Other expressions/ functions that need to access the original input buffer get wrong results.

+ This fix also sets readerIndex and writerIndex correctly for the output buffer, otherwise the consumer of the output buffer will hit issues.

closes #604

    • -0
    • +4
    /exec/java-exec/src/test/resources/functions/conv/conv.json
[maven-release-plugin] prepare release drill-1.8.0

    • -1
    • +1
    /contrib/data/tpch-sample-data/pom.xml
  1. … 12 more files in changeset.
DRILL-4800: Parallelize column reading. Read/Decode fixed width fields in parallel Decoding var length columns in parallel Use simplified decompress method for Gzip and Snappy decompression. Avoids concurrency issue with Parquet decompression. (It's also faster). Stress test Parquet read write Parallel column reader is disabled by default (may perform less well under higher concurrency)

DRILL-4852: Fix performance regression for COUNT(*) query over large JSON table

close apache/drill#576

DRILL-4857: Maintain pruning status and populate ParquetGroupScan's entries field with only the selection root if no partition pruning was done.

close apache/drill#575

DRILL-1268: Add unit test to C++ native client

Add CppUnit unit test to the C++ native client

    • -2
    • +42
    /contrib/native/client/CMakeLists.txt
    • -0
    • +67
    /contrib/native/client/cmakeModules/FindCppUnit.cmake
    • -5
    • +16
    /contrib/native/client/readme.win.txt
    • -0
    • +40
    /contrib/native/client/src/test/CMakeLists.txt
    • -0
    • +51
    /contrib/native/client/src/test/UtilsTest.cpp
    • -0
    • +39
    /contrib/native/client/src/test/main.cpp
DRILL-4726: Dynamic UDF Support

1) Configuration / parsing / options / protos

2) Zookeeper integration

3) Registration / unregistration / lazy-init

4) Unit tests

This closes #574

    • -0
    • +5
    /distribution/src/resources/sqlline.bat
  1. … 57 more files in changeset.
DRILL-4854: Fix logic error in drill-config.sh

The check of the log directory should use an OR instead of AND.

closes #572

DRILL-4853: Update C++ protobuf source files

Add support for prepared statements and metadata querying

    • -0
    • +10
    /contrib/native/client/readme.linux
    • -13
    • +7
    /contrib/native/client/scripts/fixProtodefs.sh
    • -607
    • +9515
    /contrib/native/client/src/protobuf/User.pb.cc
    • -505
    • +7594
    /contrib/native/client/src/protobuf/User.pb.h
DRILL-4452: Uses Apache Calcite Avatica driver vs Optiq driver for Drill JDBC

Drill JDBC driver uses Optiq Avatica as its basis, but this dependency has

been moved over to Calcite, for quite some time without Drill code being

updated for it.

This patch updates Avatica version to the version from Calcite

(1.4.0-drill-r19). It also refactors Drill JDBC driver to comply with the

packages and API changes in Avatica. Finally it fixes the the SQL types for

lists and structs, since Drill doesn't support java.sql.Array and

java.sql.Struct interfaces.

this closes #395

Change-Id: Ia608adf900e8708d9e6f6f58ed41e104321a9914

  1. … 8 more files in changeset.
DRILL-4420: C++ API for metadata access and prepared statements

Add support to the C++ client for metadata querying and prepared

statement requests.

Part of the metadata API, add methods to query for server capabilities.

As of now, this interface is not backed up by any RPC exchange so

the information is pretty much static, and match Drill 1.8.0

current capabilities.

    • -0
    • +123
    /contrib/native/client/src/clientlib/collectionsImpl.hpp
    • -103
    • +134
    /contrib/native/client/src/clientlib/drillClient.cpp
    • -0
    • +9
    /contrib/native/client/src/clientlib/env.h.in
    • -0
    • +406
    /contrib/native/client/src/clientlib/fieldmeta.cpp
    • -0
    • +748
    /contrib/native/client/src/clientlib/metadata.cpp
    • -0
    • +288
    /contrib/native/client/src/clientlib/metadata.hpp
  1. … 19 more files in changeset.
Edits to Drill 1.8 doc updates

Updates to docs for Drill 1.8

edit config option intro to include note about change to default drill setting locatin - /bin/drill-config.sh

update to partition pruning intro to include refresh command for metadata cache file

DRILL-4833: Insert exchanges on the inputs of union-all such that the parent and children can be independently parallelized.

Add planner option to enable/disable distribution for union-all.

close apache/drill#566

DRILL-4836: Handle NodeExistsException in ZookeeperClient#put (race condition)

ZK Issue during Drillbit startup, possibly due to race condition.

A change made in February created a race condition if two Drillbits

attempt to create the same storage plugin node at the same time.

Revised the code to eliminate the race condition by relying on an

exception to detect that the node already exists.

closes #564

updates for Drill 1.8 release - release notes, blog, version.json

    • -0
    • +106
    /_docs/rn/003-1.8.0-rn.md
    • -0
    • +31
    /blog/_posts/2016-08-15-drill-1.8-released.md
Update partition pruning intro for 1.8 - pp on parquet metadata cache

DRILL-4795: Nested aggregate windowed query fails - IllegalStateException

close apache/drill#563

edit

DRILL-4825: Fix incorrect result issue caused by partition pruning when same table is queried multiple times with different filters in query.

1) Introduce DirPrunedEnumerableTableScan which will take file selection as part of digest.

2) When directory-based pruning happens, create instance of DirPrunedEnumerableTableScan.

edit

add HashJoin's not fully parallelized in query plan - new parameters for DRILL-4743