Je vous renvoie à
But, range queries are not supported by the GIN index; asking for all products with the SIZE variations having value in a certain number range will not utilize the index. In cases like this, it is necessary to move some fields out of JSON documents to separate columns/tables and index them traditionally, using old and proven B-trees. In these instances, Mongo is more forgiving and flexible. It allows for strategies of the kind:
，这一点在新收录的资料中也有详细论述
Владислав Уткин，这一点在新收录的资料中也有详细论述
$219.99 at Best Buy。新收录的资料是该领域的重要参考
built as workarounds (like the icomplete vertical prefix indicators)