I'm not sure if this is related, but I've hit another challenge with TIMESTAMP_NTZ columns

As soon as I calculate the statistics on a TIMESTAMP_NTZ column in a table, I can't use that column in a WHERE clause date range.

This query

-- set the variable in advance to eliminate any cast/conversion issues in WHERE clause
DECLARE OR REPLACE VARIABLE dt_begin TIMESTAMP_NTZ DEFAULT '2023-12-04T16:00';

SELECT    event.timestamp as time
        , event.FirstIP4Record as DestIP        
        , event.DomainName as DNSName
        , dt_begin
FROM dnsrequest event
WHERE 
  timestamp >= dt_begin
--   timestamp IS NOT NULL

returns

MatchError: TimestampNTZType (of class org.apache.spark.sql.types.TimestampNTZType$)
com.databricks.backend.common.rpc.SparkDriverExceptions$SQLExecutionException: scala.MatchError: TimestampNTZType (of class org.apache.spark.sql.types.TimestampNTZType$)
	at com.databricks.sql.optimizer.statsEstimation.FilterEstimation.evaluateBinary(FilterEstimation.scala:523)
	at com.databricks.sql.optimizer.statsEstimation.FilterEstimation.calculateSingleCondition(FilterEstimation.scala:400)