Extract by Expression w/ two conditions stopping when first condition is met in QGIS

Question

I have an output from ST-DBSCAN with the attribute "CLUSTER_SIZE". I'd like to extract only attributes belonging to clusters of size >= 10. If this condition cannot be met, then extract attributes with clusters of size >= 7. If this condition cannot be met, NULL can be returned.

I am working with the following code:

CASE
WHEN "CLUSTER_SIZE" >= 10 THEN TRUE
WHEN "CLUSTER_SIZE" >= 7 THEN TRUE
END

Since layers with CLUSTER_SIZE >= 10 will always also have CLUSTER_SIZE >= 7, the algorithm needs to stop if the first condition is already met. According to QGIS documentation for CASE, this code would provide exactly that:

CASE is used to evaluate a series of conditions and return a result for the first condition met. The conditions are evaluated sequentially, and if a condition is true, the evaluation stops, and the corresponding result is returned.

Yet outputs for layers with CLUSTER_SIZE >= 10 are instead filtered by CLUSTER_SIZE >= 7. It seems CASE is not stopping at the first condition.

I have also tried:

CASE
    WHEN "CLUSTER_SIZE" >= 10 THEN TRUE
    ELSE "CLUSTER_SIZE" >= 7
END

and

if("CLUSTER_SIZE" >= 10, TRUE, if( "CLUSTER_SIZE" >= 7, TRUE, FALSE))

But the problem persists. How can I write a conditional code for QGIS' "Extract by Expression" tool that returns if the first condition is met and continues only if it's not?

For clarification, the outputs I need are:

• Layers only with clusters >= 10 if they have clusters of at least this size
• Layers only with clusters >= 7 if they do not have any clusters bigger than 10

I am basically loosening my filter in case I have layers with clusters smaller than 10, but this is to be done as an exception only.

Answer 1

You need to check, whether you cluster size is equal to or exceeds 10, before selecting features. You can use maximum() for this task.

if(maximum("CLUSTER_SIZE")>=10,"CLUSTER_SIZE">=10,"CLUSTER_SIZE">=7)

does, what you're looking for. maximum("CLUSTER_SIZE")>=10 checks, whether your cluster sizes meet or exceed 10. If that is true, all clusters with a size 10+ are selected. Otherwise all clusters with 7+ features are selected.