Is mongodb still necessary to divide tables?

  mongodb, question

Mongodb has its own autosharding, so is it necessary to divide tables and add a table to pass the hundred million level?

Well, I stepped on it because such an answer is misleading.
No matter what kind of database, the cost of indexing is huge, because it means to traverse the entire table of data, how can there be little pressure? That’s why there is{background: true}Options can appropriately alleviate this situation. When creating indexes in a cluster that is under too much pressure, we suggest to use “rolling” to create indexes, which will be removed from nodes one by one to create indexes and put on line to avoid affecting the operation of online systems.
As for the problem of locks, the WT engine supports document locks (row locks) from 3.0 onwards.
When querying the index, it costs a lot, mostly because your index is not properly established. You can give specific examples to discuss.
Please give specific examples to discuss when the data exceeds 100 million.