How to create a data frame in R (without using nested for loops), which records estimations from oth-CodePudding

I want to store the distances between each point (x,y) of two 2D trajectories. So I have 4 vectors: ax, ay that has time dependent coordinates of particle a and bx, by for particle b. I want to create a data frame dist that will basically be:

dist[t1,t2] = sqrt((ax[t1]-bx[t1])**2 (ay[t1]-by[t2])**2)

Now I want to do this without using nested for loops with t1 and t2, as it will take a lot of time for tmax~1000. Is there any cool dataframe trick or any of the apply methods to achieve this?

CodePudding user response：

First, let's make the data reproducible. You should do that yourself in your next questions.

ax <- 1:3
ay <- 4:6
bx <- 3:1
by <- 6:4

R being vectorized, you don't need to include indexes in your statement.

dist <- sqrt((ax-bx)^2 (ay-by)^2)
result <- data.frame(ax, ay, bx, by, dist)
> result
  ax ay bx by     dist
1  1  4  3  6 2.828427
2  2  5  2  5 0.000000
3  3  6  1  4 2.828427

CodePudding user response：

The outer function is useful for this, you could do:

outer(seq_along(ax), seq_along(bx), 
function(t1,t2){sqrt((ax[t1]-bx[t2])**2 (ay[t1]-by[t2])**2)})

I've made a slight change to your calculation: bx[t1] became bx[t2], which I think is correct