is there a way to limit queries taking a lot of time preparing execution plan ?

A while back Michael from the doc. team told me in the break room about directed queries. I wonder if that would help.

https://my.vertica.com/docs/8.1.x/HTML/index.htm#Authoring/SQLReferenceManual/Statements/DIRECTEDQUERY/DirectedQueryStatements.htm?TocPath=SQL%20Reference%20Manual|SQL%20Statements|Directed%20Query%20Statements|_____0

Hi Tom,
thank you !
but, all queries have the same behavior but are not the same (different tables, columns, analytic functions). Thinking directed queries cannot help here.
BTW, customer is trying to prevent query execution instead to run it.

Thanks you

I don't think I understand your original question. It sounds like you're building queries and that takes a long time.

Exactly, the engine takes a lot of time (hours) to prepare the execution plan of the query. During this period, the query is not visible into the query_request table, and it is using a lot of system resources.

Maybe lower the value of the MaxOptMemMB parameter?

dbadmin=> select parameter_name, current_value, description from configuration_parameters where parameter_name = 'MaxOptMemMB';
 parameter_name | current_value |                                                                 description
----------------+---------------+----------------------------------------------------------------------------------------------------------------------------------------------
 MaxOptMemMB    | 100           | Maximum amount of memory used by the Optimizer; Increasing this value may help with 'Optimizer memory use exceeds allowed limit' errors (MB)
(1 row)

Keep in mind this change will affect all users/queries...

A better solution might be to break up the complex SQL into smaller stages that include temp tables to hold intermediary results?

do you mean setting maxmemorysize for metadata resource pool < 1GB ?
Now metadata resource pool has the default setting, memorysize=0% maxmemorysize=unlimited

Taking hours for a plan that's is odd, sounds like something Engineering should take a look.