Nope, I couldn't replicate it even with all the records (removing the rest of the fields). Is there any debbugging step I can take? Anyway of adding further logs or something to know what's going on? It's a certaing combination of filters that seems to give only 1 or 0 results (even knowing there's more). If I remove one of the &&, more of them appear. I can reproduce it with the original dataset, but not with a trimmed down (600k+ but only two fields)