duplicate_distance works?

[cascafico]

I have an import candidate dataset which contains points (addresses) sometimes close to each other less than 1 meter (typically addr:housenumber=12, addr:housenumber=12a, etc.).

No matter which duplicate_distance I try (100, 5, 1, 0.5, 0), conflator logs the same duplicates, ie:

13:41:08 Dataset points duplicate each other: 21 and 455
13:41:08 Dataset points duplicate each other: 22 and 23
13:41:08 Dataset points duplicate each other: 33 and 554
13:41:08 Dataset points duplicate each other: 35 and 216
13:41:08 Dataset points duplicate each other: 72 and 845