[BUG] Databricks 13.3 executor side broadcast failure

[tgravescs]

Describe the bug
Running 24.02 snapshot on Databricks 13.3 I ran into the following error on a custom query:

Caused by: org.apache.spark.SparkException: Unexpected build plan for Executor Side Broadcast Join: 
ColumnarToRow
+- AQEShuffleRead ebj
   +- ShuffleQueryStage 132, Statistics(sizeInBytes=7.8 MiB, rowCount=2.74E+5, ColumnStat: N/A, isRuntime=true)
      +- GpuColumnarExchange gpuhashpartitioning(cast(user_id#83458 as int), 200), ENSURE_REQUIREMENTS, [plan_id=394739]
         +- GpuProject [gpucoalesce(user_id#84619, cast(user_id#84639 as string)) AS user_id#83458, if ((gpucoalesce(is_deleted#84640, true) AND NOT (cast(user_id#84619 as int) = 117563))) gdpr-deleted else gpucoalesce(email_address#84631, email_address#84646) AS email_address#83462]
            +- GpuRowToColumnar targetsize(268435456)
               +- *(92) BroadcastHashJoin [org_name#84630], [org_name#84647], LeftOuter, BuildRight, false
                  :- GpuColumnarToRow false
                  :  +- GpuFilter gpuisnotnull(gpucoalesce(user_id#84619, cast(user_id#84639 as string)))
                  :     +- GpuShuffledHashJoin [user_id#84619], [cast(user_id#84639 as string)], FullOuter, GpuBuildRight, false
                  :        :- GpuProject [user_id#84619, org_name#84630, email_address#84631]
                  :        :  +- GpuCoalesceBatches targetsize(268435456)

First part of stack trace:

 at org.apache.spark.sql.execution.joins.ExecutorBroadcast$.getShuffleIdFromPlan(ExecutorBroadcast.scala:189)
        at org.apache.spark.sql.execution.joins.BroadcastHashJoinExec.executorBroadcast$lzycompute(BroadcastHashJoinExec.scala:73)
        at org.apache.spark.sql.execution.joins.BroadcastHashJoinExec.executorBroadcast(BroadcastHashJoinExec.scala:71)
        at org.apache.spark.sql.execution.joins.BroadcastHashJoinExec.prepareBroadcast(BroadcastHashJoinExec.scala:284)
        at org.apache.spark.sql.execution.joins.BroadcastHashJoinExec.prepareRelation(BroadcastHashJoinExec.scala:304)
        at org.apache.spark.sql.execution.joins.HashJoin.codegenOuter(HashJoin.scala:451)
        at org.apache.spark.sql.execution.joins.HashJoin.codegenOuter$(HashJoin.scala:450)
        at org.apache.spark.sql.execution.joins.BroadcastHashJoinExec.codegenOuter(BroadcastHashJoinExec.scala:53)
        at org.apache.spark.sql.execution.joins.HashJoin.doConsume(HashJoin.scala:359)
        at org.apache.spark.sql.execution.joins.HashJoin.doConsume$(HashJoin.scala:356)
        at org.apache.spark.sql.execution.joins.BroadcastHashJoinExec.doConsume(BroadcastHashJoinExec.scala:53)
        at org.apache.spark.sql.execution.CodegenSupport.consume(WholeStageCodegenExec.scala:199)
        at org.apache.spark.sql.execution.CodegenSupport.consume$(WholeStageCodegenExec.scala:154)
        at org.apache.spark.sql.execution.InputAdapter.consume(WholeStageCodegenExec.scala:503)
        at org.apache.spark.sql.execution.InputRDDCodegen.doProduce(WholeStageCodegenExec.scala:490)