Share a way to speed up retrieval

I've just discovered by accident, and I've just realized that this simple way to do it is just now that I'm thinking... The simple thing is, if your code needs to compute some parameters with library functions, and you actually only need the most recent data in the result set, then you don't have to directly input the raw K-line data into the function, you just have to input the most recent N-line data, like the moving average in js.TA.MA(records.slice ((-N))), the second moving average data that needs to be countered isTA.MA(records.slice ((-N-1)) ). The principle is simple: if the length of the K-line data is Length, then the library function needs to compute the Length-N+1 indicator, so the shorter the data length, the fewer the computations, the faster the speed. The indicator is only related to the most recent N data, so the previous data can be safely ignored. Of course, N is not absolute, some indicators are calculated based on the most recent N+1 data, and even some indicators require the previous value of the indicator to be used, so you need to keep one or two hundred data. This is a relatively small discovery, so if you already know, please ignore me...