Possibility to disable background parsing job

[viperh]

Is your feature request related to a problem? Please describe.
In a specific way yes
Describe the solution you'd like
A settings option to be able to disable the background parsing job

Describe alternatives you've considered
To improve the performance of the background parsing job

Additional context
The reason I request this is that on my PC, DBeaver is very slow when making queries with a good amount of data.
I have tested this on other computers from my colleagues as well, and the performance issue persists.
I've managed to track this down and speed it up by manually cancelling the background parsing job.
On average, I've waited around 2,5 min for 1 query. By disabling the job, it reduced to a few seconds ~4 seconds give or take.

[ged-yuko]

An option to completely disable background parsing is Enable semantic analysis.

You can also try to disable just the Read database table/column for semantic analysis option. This way background analysis will be providing some advanced highlighting but won't use any information about the database structure, and so it won't be accessing database connection at all.

Note that it is not really related with query execution. It does some analysis and loads metadata information about the database schema while analyzing and validating query, i breifly described the whole procedure here.
So the only reason for the query execution to conflict with background analysis i may think of is during editor initialization at the moment of connection establishment, when the metadata cache is not populated yet.

It would be nice if you can repeat your test using following scenario with both mentioned options enabled:

• open your sql script containing the query in question;
• place the editing cursor at the query you want to measure the execution time of;
• ensure that the connection is established and wait for all the background activities to finish by looking at the Background Tasks view, until it'll be empty;
• use Execute SQL query command and measure its exclusive execution time without other activities, there should be only query-execution-related activities going on (it also requests some database metadata by the way, and uses the same cache as analysis).

Then repeat it with Enable semantic analysis enabled and Read database table/column for semantic analysis disabled.
Then repeat it with Enable semantic analysis disabled.

Also it might be that your other configuration options discourage "separate connection for metadata read", which effectively prevents parallelization of metadata reading and queries execution, because forces DBeaver to use one and the same database connection for all the operations. This i cannot guess. "Background analysis" called "background" for a reason after all.
There is such an option by the way - to force or prevent separate metadata connections. It's just not always works depending on the database configuration.

You can use Query Manager to track which queries executed with which connection and for how long. It might even be that your database or the network connection struggles when loading both metadata and query results simultaneously.

PS I should explicitly say, that the issue is apparently not about the background analysis performance but about the exagreggated timing of the query execution. And the issue description completely lack the scenario in which it is being observed. So at the moment no meaningful details really presented to allow its solution or explaination of actual reasons for.

[viperh]

epic. worked.