为什么 RGUI 会自动关闭？

Question

该代码旨在分析沿每条折线的点位置。

它在 ttl_ln 小时有效。 但是当整个折线要素加载完毕（超过 300,000 条折线）时，Rgui 会自动关闭。

我通过添加 bool test_tag 进行了一些测试。当 test_tag 为 true 时，Rgui 关闭发生，当 false 时，则可以仅使用 subset() 并在循环中重复矩阵创建来处理 300,000。所以可能不是因为使用了 subset() 或 matrix()。

使用 Rcpp 中定义的 scale_to_grid() 时会关闭。

shp_file = 'some_ply.shp'
shp_1 = sf::st_read(shp_file)
shp_2 = sf::st_coordinates(shp_1)
ttl_ln = nrow(shp_1)
for (ln in c(1:ttl_ln)){
    
    shp_3 = subset(shp_2, shp_2[,3] == ln)

    # field 1 is coordinate X
    # field 2 is coordinate Y
    # field 3 is Line ID created by sf package
    
    em_mat = matrix(0, nrow = 200, ncol=200)
    
    if(test_tag){

        # manipulations on points are followed
        # a groupe of functions defined in a Rcpp file
        # are used. They can be successfully compiled.

        new_mat = scale_to_grid(shp_3, em_mat, t1, t2, t3, t4)
        # t1, t2, t3, t4 are numbers with double type in Rcpp file. 
    }
    
     
}

以下是 Rcpp 中的代码。

// [[Rcpp::export]]
NumericMatrix scale_to_grid(NumericMatrix shp, 
                            NumericMatrix mat, 
                            double o_x, double o_y,
                            double u_x, double u_y){

    // mat : the matrix of grid

    int ptr;
    int nbr  = shp.nrow();
    int ind_x, ind_y;       
    double slope, new_x, new_y;


    // go through each point, expect the last one (i.e. ptr < nbr -1)
    for(ptr=0;ptr<nbr-1;ptr++){
        ind_x = trunc((shp(ptr,0)-o_x)/u_x);
        ind_y = trunc((shp(ptr,1)-o_y)/u_y);        
    
        mat(ind_x, ind_y)++;
    
        if(abs(shp(ptr+1,0) - shp(ptr,0)) > u_x){
            // the difference in x value is greater than the grid unit width
            // there must be grid to fill with value '1'
            slope = (shp(ptr+1,1) - shp(ptr,1)) / (shp(ptr+1,0) - shp(ptr,0));
        
            // initialise
            new_x = shp(ptr,0); 
            new_y = shp(ptr,1); 
        
            while(abs(new_x - shp(ptr+1,0)) > u_x){
            
                if(shp(ptr+1,0) - shp(ptr,0) > 0){
                    new_x = new_x + u_x;
                    new_y = new_y + u_x * slope;
                }else{
                    new_x = new_x - u_x;
                    new_y = new_y - u_x * slope;                    
                }
            
            ind_x = trunc((new_x-o_x)/u_x);
            ind_y = trunc((new_y-o_y)/u_y);     
            mat(ind_x, ind_y)++;                
            }

        }

    }
return mat;}