Running queries using CQL containing OR subexpression sometimes yield duplicated results. We’ve attached a sample data file and converter configs for reproducing this problem. Ingest test data
Now we have two schemas containing exactly the same data. Both datasets have an attribute I containing numeric values. seq_with_index has the integer attribute I indexed while seq_no_index do not have attribute indexes.
Reproduce problem 1 Query the dataset using CQL (I > 5 AND I < 10) OR I > 95 on schema with attribute index yield duplicated results (#96 to #99):
Those duplicated results goes away when we run the same query on schema without attribute index:
Reproduce problem 2 When the OR subexpression contains several range queries with overlapping ranges, running the query on indexed attribute would produce duplicated results:
Those duplicated results goes away when we run the same query on schema without attribute index:
Expected behavior No duplicated results should be retrieved when running queries mentioned above. |