I\'ve got millions of items ordered by a precomputed score. Each item has many boolean attributes. Let says that there is about ten thousand possible attributes totally, each
This is exactly what Mongo can deal with. The fact that your attributes are boolean type helps here. A possible schema is listed below:
[
{
true_tags:[attr1, attr2, attr3, ...],
false_tags: [attr4, attr5, attr6, ...]
},
]
Then we can index on true_tags and false_tags. And it should be efficient to search with $in, $all, ... query operators.