Signed-off-by: sperlingxx <lovedreamf@gmail.com>

* Depend on the cuDF v0.18

Change rapids brannch-0.4 to depend on cuDF v0.18 release jars

Prepare for the for the rapids v0.4.0 release

Signed-off-by: Tim Liu <timl@nvidia.com>

* cudf 0.17-SNAPSHOT to 0.17

…IDIA#1808)

* mortgage support multiple dataset formats

change mortgage sample class to support dataset formats csv/orc/parquet

Signed-off-by: Tim Liu <timl@nvidia.com>

* Update

1, copyright 2021
2, throw an error if there are more than 5 arguments
3, match-case optimize

Signed-off-by: Tim Liu <timl@nvidia.com>

* Update

1, print some helpful info for the input arguments
2, exit instead of exeption, when arguments are wrongly set

* fix typo

* Fix Nothing value in 'case _ =>'

* update

Signed-off-by: Jason Lowe <jlowe@nvidia.com>

Fix merge conflict with branch-0.4

* Spark 3.0.2 shim no longer a snapshot shim

Signed-off-by: Jason Lowe <jlowe@nvidia.com>

* Remove 3.0.2-SNAPSHOT support

[auto-merge] branch-0.4 to branch-0.5 [skip ci] [bot]

Signed-off-by: Thomas Graves <tgraves@nvidia.com>

Signed-off-by: Gera Shegalov <gera@apache.org>

Add a shim provider for Spark 3.2.0 development branch. Closes NVIDIA#1490
- fix overflows in aggregate buffer for GpuSum by wiring the explicit output column type
- unit tests for the new shim
- consolidate version profiles in the parent pom

* Cleanup unused Jenkins files and scripts

NVIDIA#1568

Move Databricks scripts to GitLab so we can use the common scripts for the nightly build job and integration tests job

Remove unused Dockerfiles

Signed-off-by: Tim Liu <timl@nvidia.com>

* rm Dockerfile.integration.ubuntu16

* Restore Databricks nightly scripts

Signed-off-by: Tim Liu <timl@nvidia.com>

* Spark 3.1.1 shim no longer a snapshot shim

Signed-off-by: Jason Lowe <jlowe@nvidia.com>

* Remove 3.1.0, 3.1.0-SNAPSHOT, and 3.1.1-SNAPSHOT support

* Remove obsolete comment

* Update to note support for 3.0.2

Signed-off-by: Sameer Raheja <sraheja@nvidia.com>

* Update FAQ to reflect 3.0.2 and 3.1.1 support

Signed-off-by: Sameer Raheja <sraheja@nvidia.com>

In the 'Map' of dataset-format, the function of 'Run.csv()/Run.orc()/Run.parquet' will be executed one by one, then it causes the dataset format error, because the dataset format in the current test is 'parquet'

Change 'Run.csv()/Run.orc()/Run.parquet' into the lambda expressions, to avoid running the 'Run.xxx()' functions in the dataFrameFormatMap

Signed-off-by: Tim Liu <timl@nvidia.com>

Signed-off-by: Robert (Bobby) Evans <bobby@apache.org>

Port apache/spark#31320 to close NVIDIA#1735

Signed-off-by: Gera Shegalov <gera@apache.org>

Fix merge conflict with branch-0.4

* Update changelog for 0.4

Signed-off-by: Sameer Raheja <sraheja@nvidia.com>

* Update generate-changelog script

Signed-off-by: Sameer Raheja <sraheja@nvidia.com>

[auto-merge] branch-0.4 to branch-0.5 [skip ci] [bot]

* Refactor join code to reduce duplicated code

Signed-off-by: Jason Lowe <jlowe@nvidia.com>

* Move nodeName override to base class

* Add shim for Spark 3.0.3

Signed-off-by: Jason Lowe <jlowe@nvidia.com>

* Add premerge testing for Spark 3.0.2 and Spark 3.0.3

* Fix Part Suite Tests

Signed-off-by: Robert (Bobby) Evans <bobby@apache.org>

* Addressed review comments

* Add shim for Spark 3.1.2

Signed-off-by: Jason Lowe <jlowe@nvidia.com>

* Add Spark 3.1.2 to premerge testing

* fix shuffle manager doc on ucx library path

Signed-off-by: Rong Ou <rong.ou@gmail.com>

* remove ld library path line

Signed-off-by: Rong Ou <rong.ou@gmail.com>

…VIDIA#1871)

Signed-off-by: Jason Lowe <jlowe@nvidia.com>

Signed-off-by: Robert (Bobby) Evans <bobby@apache.org>

Signed-off-by: Niranjan Artal <nartal@nvidia.com>

Fix merge conflict with branch-0.4

Signed-off-by: sperlingxx <lovedreamf@gmail.com>