Thanks a lot, it seems that disabling the extra_cat_name(categories) udf fixes the duplicates, I however don't understant why, if you have an idea