[FEA] Java bindings for fixed-point Decimal type

[sperlingxx]

Is your feature request related to a problem? Please describe.
cuDF Java API should be able to leverage new cuDF feature, fixed-point decimal type, in order to support decimal type in spark-rapids plugin (issue link).

Describe the solution you'd like
Initial Tasks

• Add class DataType, and add Decimal32/64 into DType collection ([REVIEW] Initial work for decimal type in Java/JNI [skip ci] #6514)
  • Add class NativeDataType (which manages the handle of native data_type class)
  • Package-level adaption for Dtype -> DataType transformation (check all references of DType)
  • Scalar (scalar JNI, scalar building functions for decimalType)

Other Tasks

• ColumnVectors ([REVIEW] Support fixed-point decimal for HostColumnVector [skip ci] #6609)
• BinaryOperations with decimal data
• JCudfSerialization
• Schema
• TestUtils for decimal types

Additional context
Native Prerequisites

• (merged) Move fixed_point::scale to data_type ([REVIEW] fixed_point Column Optimization (store scale in data_type) [skip ci] #5861)
• (merged) Enable fixed_point binary operations ([REVIEW] Enable fixed_point binary operations #6528 [REVIEW] Remove fixed_point precise round #6544)
• (merged) Read/write parquet/orc with fixed-point decimal data ([REVIEW] Enable strict_decimal_types in parquet reading #6969 )
• (merged) Type casting between fixed-point decimal and other numeric types ([FEA] type casting between fixed-point decimal and other numeric types #6486)
• Support HALF_UP and HALF_EVEN rounding to a set number of decimal places ([FEA] Support HALF_UP and HALF_EVEN rounding to a set number of decimal places #3790)
• Unary operators for fixed-point decimals
• Aggregation operators for fixed-point decimals

[kkraus14]

Pushing this to 0.18

[github-actions]

This issue has been labeled inactive-90d due to no recent activity in the past 90 days. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed.

[revans2]

@sperlingxx how much is left for this to be done? I think it is almost all complete

[sperlingxx]

@sperlingxx how much is left for this to be done? I think it is almost all complete

Yes, it is almost complete except some cuDF native tasks. Shall we close this issue?

[beckernick]

Are the outstanding tasks mentioned in this issue covered by any other issues? If so, perhaps this can be closed.

[harrism]

@sameerz @sperlingxx can you respond to @beckernick 's question above?

[sameerz]

The ORC read / write for decimal in cuio is done in per #7126 and can be followed up on with the JNI binding work. We need to create issues (if we do not already have them) for unary operators for fixed-point decimals and aggregation operators for fixed-point decimals.

The RAPIDS Spark tasks related to this issue are tracked in NVIDIA/spark-rapids#42