rapids

The use of an mr parameter in inplace_bitmask_and, which calls inplace_bitmask_binop, is a little misleading. The allocations there are always temporary and are not part of the return value. It is only used for a few temporary arrays/scalars: https://github.com/rapidsai/cudf/blob/1f8a03e69704562dfac38de40b7172650280c6ea/cpp/include/cudf/detail/null_mask.cuh#L169-L171

It should be possible

We no longer need to control the number of concurrent kernels, since now we control the number of concurrent tasks

rapidsai/cudf#10618 made a change to pandas and as a part of that discussion I realized that we are not testing

• What happens if there are duplicate columns. I think Spark is going to return an error when doing schema discovery, but I am not sure what happens when we pass in a schema and include column names. This is even more interesting if we read a column that i

I'm submitting a

• [x ] bug report.

Current Behaviour:

After #249
Trying to run tests with pytest tests/rdf_tests/test_rdf_basic.py -k test_rdf_runner -s, you get a report file with all the tests run.
Some tests return errors, for example:

{
    "Basic - Term 7": {
        "input": "basic/data-4.ttl",
        "query": "basic/term-7.rq",
        "error": "Expected {Sele

I was just thinking making sure that it works as a valid execution policy. So simply launch a Thrust algorithm with device data and make sure the result is sound.

I was going to say "just do what we do for our current tests", but apparently that's nothing!

Originally posted by @jrhemstad in rapidsai/rmm#1009 (comment)

In trying to write tests for #189, I'm finding very difficult to add columns to existing tests, as in some cases like the all_types table, the table is defined in a separate file than the tests and multiple tests try to write to the same table.

Additionally, our test suite doesn't prove that the data that are uploaded are the same as the data downloaded for all types.

We should consider m

As seen with gumdropsteve/turbo-telegram@3e2f3b3, the data for the first half of 2016 can be downloaded & preprocessed just like that of 2015. Is there any other data in the effective range? I.e. is pre-2015 data recorded the same?

If so, let's add it.