Very big tables - Filter acceleration

[georgmaisser]

Very big tables produce problems because of the way filters are generated. Filter generation needs to be made less expensive and async to optimize performance and usability

We need a combination of these possibilities:

• [x] "Select distinct" calls for every affected filter column
• [ ] Filter Creation in chunks (with limit)
• [ ] Lazy loaded filter options, when filter is not already created
• [ ] Set Ad hoc task to recreate filter after every cache purging which invalidates the filter

[georgmaisser]

Because of the success of the first task, we don't need the rest for the moment.