[FEA] Regular expressions - support line anchors in choice

[NVnavkumar]

Is your feature request related to a problem? Please describe.
I wish the RAPIDS Accelerator for Apache Spark supported line anchors in Regular Expression choice (|) expressions.

Describe the solution you'd like

scala> spark.conf.set("spark.rapids.sql.enabled", "false")
scala> val df = spark.range(0, 1000).selectExpr("CAST(id AS STRING) as id")

scala> df.createOrReplaceTempView("table")

scala> spark.sql("select concat(a.id, '/', b.id, '\n') as id from table as a right join  table as b on a.id = b.id").write.mode("overwrite").parquet("/tmp/test-re-data-20221021")

scala> spark.conf.set("spark.rapids.sql.enabled", "true")

scala> spark.read.parquet("/tmp/test-re-data-20221021").selectExpr("regexp_extract(id, '\\\\d+(6$|7$)', 1)").collect()
22/10/21 17:40:44 WARN GpuOverrides:
!Exec <ProjectExec> cannot run on GPU because not all expressions can be replaced
  @Expression <Alias> regexp_extract(id#103, \d+(6$|7$), 1) AS regexp_extract(id, \d+(6$|7$), 1)#105 could run on GPU
    !Expression <RegExpExtract> regexp_extract(id#103, \d+(6$|7$), 1) cannot run on GPU because cuDF does not support terms ending with line anchors on one side of a choice near index 4; regex group count is 0, but the specified group index is 1
      @Expression <AttributeReference> id#103 could run on GPU
      @Expression <Literal> \d+(6$|7$) could run on GPU
      @Expression <Literal> 1 could run on GPU

res19: Array[org.apache.spark.sql.Row] = Array([], [], [], [6], [7], [], [], [], [6], [7], [], [], [], [], [], [], [], [], [6], [7], [], [], [], [], [], [], [], [], [6], [7], [], [], [], [], [], [], [], [], [6], [], [], [], [], [], [], [], [6], [7], [], [], [], [], [], [], [], [7], [], [], [], [], [], [], [], [], [6], [7], [], [], [], [], [], [], [], [], [], [], [6], [7], [], [], [], [], [], [], [], [6], [], [], [], [], [], [], [], [], [6], [7], [], [], [], [], [], [], [], [7], [], [], [], [], [], [], [], [], [6], [7], [], [], [], [], [], [], [6], [7], [], [], [], [], [], [], [], [], [6], [7], [], [], [], [], [], [], [6], [7], [], [], [], [], [], [], [], [], [6], [7], [], [], [], [], [6], [7], [], [], [], [], [], [], [], [6], [7], [], [], [], [], [6], [7], [], ...

I would like the RegExpExtract to run on the GPU.

[NVnavkumar]

Filed rapidsai/cudf#11979 against cuDF.

This would give us the ability to pass arbitrary line terminators in and support all scenarios involving the line anchor $

[NVnavkumar]

Also, fixing this will probably involve ultimately involve using cuDF multiline mode, which means the handling of $ will require major updates.

[NVnavkumar]

Also, fixing this will probably involve ultimately involve using cuDF multiline mode, which means the handling of $ will require major updates.

Filed #7090 to track the updates to handle $.