Elasticsearch is reasonable about how to design index and type.

now there are several completely different statistical data structures for a city, which are obtained from parsing tables. If you put dozens of type, under an index in the mapping of ES, there are probably dozens of type, under different Index or all under one Index, and the amount of data is basically about one million.


since it is for a city, of course, put an index, dozens of type does not matter, if you are worried about the amount of data is too large, this is unnecessary, because the first million data is not large, and then ES will store these data in different shard shards (shard can be configured by itself).

Menu