Robert (Bobby) Evans

[FEA] first and last as hash based aggregates

4

**Is your feature request related to a problem? Please describe.** In a recent customer query I found that it was rather slow because they were doing a first aggregation along...

[FEA] parquet gpuComputePageSizes takes a very long time to compute

3

**Is your feature request related to a problem? Please describe.** I am not an expert on decoding the parquet data. So this might just be a question or perhaps it...

This depends on https://github.com/rapidsai/cudf/pull/11556

[BUG] pmod for large negative decimal values can overflow.

2

**Describe the bug** `pmod` when done with decimal values on a precision of 38 can produce incorrect answers. if the numbers are negative and large. **Steps/Code to reproduce bug** a=Decimal('-9417536006095259414705321248.3563971038')...

**Describe the bug** This reduces the quality of the tests. I have a patch for this. ```diff diff --git a/integration_tests/src/main/python/data_gen.py b/integration_tests/src/main/python/data_gen.py index ac3c812a5..4f864691a 100644 --- a/integration_tests/src/main/python/data_gen.py +++ b/integration_tests/src/main/python/data_gen.py @@ -215,17...

[BUG] 0 in some cases for decimal being cast to a string returns different results.

1

**Describe the bug** This is related to https://github.com/NVIDIA/spark-rapids/issues/6337 after applying the patch we have two tests that fail. ``` FAILED ../../src/main/python/cast_test.py::test_cast_map_to_string[true-Map(Decimal(not_null)(7,7),Decimal(12,2))] FAILED ../../src/main/python/cast_test.py::test_cast_map_to_string[false-Map(Decimal(not_null)(7,7),Decimal(12,2))] ``` It appears that they are failing...

**Describe the bug** This is after https://github.com/NVIDIA/spark-rapids/issues/6337 After applying the patch I got two new failures. ``` FAILED ../../src/main/python/hash_aggregate_test.py::test_hash_grpby_sum_full_decimal[{'spark.rapids.sql.variableFloatAgg.enabled': 'true', 'spark.rapids.sql.hasNans': 'false', 'spark.rapids.sql.castStringToFloat.enabled': 'true', 'spark.rapids.sql.batchSizeBytes': '250'}-[('a', RepeatSeq(Short)), ('b', Decimal(38,0)), ('c',...

**Describe the bug** This is related to https://github.com/NVIDIA/spark-rapids/issues/6337 After applying the patch I get failures for ``` FAILED ../../src/main/python/arithmetic_ops_test.py::test_decimal_multiplication_mixed_no_overflow_guarantees[Decimal(10,3)-DecimalType(27,7)-Decimal(15,3)-DecimalType(30,10)][IGNORE_ORDER] FAILED ../../src/main/python/arithmetic_ops_test.py::test_decimal_multiplication_mixed_no_overflow_guarantees[Decimal(3,-3)-DecimalType(20,-3)-Decimal(10,2)-DecimalType(28,9)][IGNORE_ORDER] FAILED ../../src/main/python/arithmetic_ops_test.py::test_decimal_multiplication_mixed_no_overflow_guarantees[Decimal(3,-3)-DecimalType(20,-3)-Decimal(15,3)-DecimalType(30,10)][IGNORE_ORDER] FAILED ../../src/main/python/arithmetic_ops_test.py::test_decimal_multiplication_mixed_no_overflow_guarantees[Decimal(10,3)-DecimalType(27,7)-Decimal(10,2)-DecimalType(28,9)][IGNORE_ORDER] ``` I captured some results...

[FEA] fix issues so we can remove hasNans config

36

**Is your feature request related to a problem? Please describe.** Most users do not have NaNs in their data, but a lot of users have floating point and end up...

[FEA] look at ways to use replace_nans and replace_nulls from cudf

2

As a part of the discussion at https://github.com/NVIDIA/spark-rapids/issues/6164#issuecomment-1210095670 we saw that there is a `replace_nans` function in cudf, along with a `replace_nulls` one. It might be good for us to...