Encountered a technical problem
For monitoring the driving paths of nearly 10,000 vehicles (or even more), when I was designing the database (mongdo) table (collection), I had 2 options
1. By day Divided into tables, all vehicle trajectories for each day are placed in one table
2. Divided into tables according to vehicles, each license plate corresponds to a table
What do you think is reasonable?
高洛峰2017-05-02 09:28:33
Your demand is a very common one, usually called Time Series data modeling.
1. It is not recommended to divide collections into different collections by day/car;
2. If you have a particularly large number of documents, you can consider Sharding.
Because you reveal less information, if you get the car’s position every minute, you might as well try the following method:
{ plate : String,
brand : String,
color : String,
timestamp_hour : ISODate,
track : [{0 : [Longitude, latitude]},
{1 : [xxx, xxx]}
...
{59 : [xxx, xxx]}]
}
The general idea is to try to make one document contain as much information as possible to improve query efficiency.
For reference.
Love MongoDB! Have Fun!
MongoDB Online Lecture Series 19- MongoDB 10 steps to build a single view
It’s April 19th, please click:>---<
我想大声告诉你2017-05-02 09:28:33
You need to consider the specific business scenario, how to display the data, and the display dimensions.