Extend LEAD/LAG to work with non-fixed-width types

[mythrocks]

This commit extends the LEAD()/LAG() offset window functions introduced in #6277 to work with more than just fixed-width data types. The functions are defined as follows:

1. LEAD(N): Returns the row from the input column, at the specified offset past the
   current row. If the offset crosses the grouping boundary or column boundary for
   a given row, a "default" value is returned if available. The "default" value is
   null, by default.
2. LAG(N): returns the row from the input column at the specified offset preceding
   the current row. If the offset crosses the grouping boundary or column boundary for
   a given row, a "default" value is returned if available. The "default" value is
   null, by default.

As an illustration, consider the following example input array input column, with two groups (G1 and G2):

[ [1,1,1], [2], [3,3,3], [4,4,4,4],  [66,66], [], [88,88], [99] ]
  <--------------G1-------------->   <------------G2----------->

LEAD(col, 1) yields:

[ [2], [3,3,3], [4,4,4,4], ∅, [], [88,88], [99], ∅ ]

LAG(input_col, 2) yields:

[ ∅, ∅, [2], [3,3,3], ∅, ∅, [], [88,88] ]

If a defaults column is specified with contents:

[ [999], [999], [999], [999], [999], [999], [999], [999] ]

then, LEAD(col, 1, defaults) yields:

[ [2], [3,3,3], [4,4,4,4], [999], [], [88,88], [99], [999] ]

Note that in the cases where the offset (1) would cross the column/group boundary (i.e. indices 3 and 7), the corresponding entry from the defaults column is returned instead of ∅ (i.e. null).

[mythrocks]

Looks to be a transient test failure:

11:21:25 FATAL: Non-zero status 255: [nvidia-smi, -L]
11:21:25 java.lang.RuntimeException: Non-zero status 255: [nvidia-smi, -L]
11:21:25 	at com.gpuopenanalytics.jenkins.remotedocker.config.NvidiaGpuDevicesConfigItem.executeWithOutput(NvidiaGpuDevicesConfigItem.java:130)
11:21:25 	at com.gpuopenanalytics.jenkins.remotedocker.config.NvidiaGpuDevicesConfigItem.addCreateArgs(NvidiaGpuDevicesConfigItem.java:83)
11:21:25 	at com.gpuopenanalytics.jenkins.remotedocker.job.DockerImageConfiguration.lambda$addCreateArgs$0(DockerImageConfiguration.java:140)
11:21:25 	at java.util.ArrayList$ArrayListSpliterator.forEachRemaining(ArrayList.java:1384)
11:21:25 	at java.util.stream.ReferencePipeline$Head.forEach(ReferencePipeline.java:647)
11:21:25 	at com.gpuopenanalytics.jenkins.remotedocker.job.DockerImageConfiguration.addCreateArgs(DockerImageConfiguration.java:140)
11:21:25 	at com.gpuopenanalytics.jenkins.remotedocker.DockerState.getlaunchArgs(DockerState.java:275)
11:21:25 	at com.gpuopenanalytics.jenkins.remotedocker.DockerState.launchContainer(DockerState.java:285)
11:21:25 	at com.gpuopenanalytics.jenkins.remotedocker.DockerState.launchContainers(DockerState.java:199)
11:21:25 	at com.gpuopenanalytics.jenkins.remotedocker.RemoteDockerBuildWrapper.setUp(RemoteDockerBuildWrapper.java:175)
11:21:25 	at hudson.model.Build$BuildExecution.doRun(Build.java:157)
11:21:25 	at hudson.model.AbstractBuild$AbstractBuildExecution.run(AbstractBuild.java:513)
11:21:25 	at hudson.model.Run.execute(Run.java:1907)
11:21:25 	at hudson.matrix.MatrixRun.run(MatrixRun.java:153)
11:21:25 	at hudson.model.ResourceController.execute(ResourceController.java:97)
11:21:25 	at hudson.model.Executor.run(Executor.java:429)
11:21:25 Recording test results

[codecov]

Codecov Report

Merging #8062 (16eb271) into branch-0.20 (51336df) will increase coverage by 0.01%.
The diff coverage is n/a.

@@               Coverage Diff               @@
##           branch-0.20    #8062      +/-   ##
===============================================
+ Coverage        82.88%   82.90%   +0.01%     
===============================================
  Files              103      103              
  Lines            17668    17877     +209     
===============================================
+ Hits             14645    14821     +176     
- Misses            3023     3056      +33     

Impacted Files	Coverage Δ
python/cudf/cudf/core/abc.py	91.66% <ø> (+0.17%)	⬆️
python/cudf/cudf/core/algorithms.py	82.35% <ø> (ø)
python/cudf/cudf/core/column/__init__.py	100.00% <ø> (ø)
python/cudf/cudf/core/column/categorical.py	92.37% <ø> (+0.13%)	⬆️
python/cudf/cudf/core/column/column.py	88.20% <ø> (-0.44%)	⬇️
python/cudf/cudf/core/column/datetime.py	88.03% <ø> (-1.88%)	⬇️
python/cudf/cudf/core/column/decimal.py	90.29% <ø> (-2.64%)	⬇️
python/cudf/cudf/core/column/interval.py	91.66% <ø> (+0.55%)	⬆️
python/cudf/cudf/core/column/lists.py	86.98% <ø> (-0.43%)	⬇️
python/cudf/cudf/core/column/numerical.py	94.72% <ø> (+0.29%)	⬆️
... and 44 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 27ae8c1...16eb271. Read the comment docs.

[ttnghia]

Rerun tests.

[mythrocks]

Rerun tests

[mythrocks]

Thanks for the reviews, chaps! I'll wait for CI to pass before checking this in.

[harrism]

@gpucibot merge