How much users have to vote for it, that this will be reconsidered?

Well done Atlassian! Your users start giving up on this... May be that is your strategy?

This needs to be reconsidered. We already use Crowd and still have a major problem sorting the users. 

This is not ideal...

I used the following SQL query to get a list of usernames from a given directory (in my case the Jira Internal Directory):

SELECT lower_user_name FROM cwd_user INNER JOIN cwd_directory ON directory_id = cwd_directory.id WHERE directory_name='Jira Internal Directory'; 

A safer option would probably be to access this information through the Jira Java API and make it a plugin out of it. I didn't get as far as trying it, but the ApplicationUser class contains the getDirectoryId, returning the directoryId the user is in:

https://docs.atlassian.com/software/jira/docs/api/8.22.0/com/atlassian/jira/user/ApplicationUser.html

From there you might be able to use the embedded Crowd API's CrowdDirectoryService to find the Directory object based on that id. The Directory object has a getName method that would give you the name.

This truly is a pain for customers with multiple directories.  I have found a workaround for this problem if you also have Asset Insight AND have configured the Jira integration to create a "Jira" schema within Asset Insight (performance of this integration is another topic altogether).  This "Jira" schema will allow you to quickly sort or query by Directory within the data collected by the integration from Jira's User Management system.

Such a critical simple feature that many admins could benefit from. In our environment IT manages the Crowd service so as an admin of a specific Jira instance I have to run SQL queries to see which users are in which directories as I do not have access to Crowd. All I want is what people are asking for a simple dropdown filter for directory.

I agree. It's obviously a political decision. We see no use for Crowd in our organization though.

It is annoying but I can see doing this to encourage people to use Crowd. Not very helpful that.
I've used ScriptRunner scripts to find this information in the past, but that requires some knowledge of the Java API for Jira

Why is this "not being considered"? It's so obviously useful, and seems like such a simple thing to implement!

Crowd is not the solution. I am using Crowd to read multiple AD's. 
But in Jira I also read the AD to let users just Log In. Those thousands shouldn't be managed in Crowd because of no use.

Meaning, we miss a simple filter option in Jira to filter on Directories.

And yes, we know that there might be an ADD-ON for that...... but com'on.  Just at a simple filter option at the top and we are very happy !