count(*) over table in offline mode

We don't support agg over table in offline mode, only online mode supports.
So if user wants to get offline data count, it's hard to do. Now user can run it on spark-shell/pyspark.
But if user can't prepare the spark env, it's better to enable sparksql in openmldb.

pass --spark_conf with option `spark.openmldb.sparksql`

P.S. user should know that not all openmldb sql can be run on spark engine.

Look for the offline mode execution path and the SQL parser. The issue mentions enabling sparksql via a spark.conf flag `spark.openmldb.sparksql`. Find where offline queries are routed and how to conditionally use Spark's SQL engine for aggregate functions like count(*). Check test files for offline query execution to understand the expected behavior.

Escrito por el modelo de indexación a partir del texto del issue.