We can create a generic projector that would load or create the document, and we will keep it. Handlers could update the document content, and we flush it either all the time or periodically. Flushing would do a full replace, which would work faster for databases like Elasticsearch, but I am not sure about Mongo. Any key-value store would work, too, without requiring to load the document first.
We can create a generic projector that would load or create the document, and we will keep it. Handlers could update the document content, and we flush it either all the time or periodically. Flushing would do a full replace, which would work faster for databases like Elasticsearch, but I am not sure about Mongo. Any key-value store would work, too, without requiring to load the document first.