如何获得不相交的积分，r

Question

我在坐标系中有两组：A和B。B中的所有点都在A中。因此，我想选择不在B中的A中的点。 我正在使用SF库，但我有点困惑。

在这里，我显示了一个可再现的Exameple。考虑下一个数据。

fname <- system.file("shape/nc.shp", package="sf")
nc <- st_read(fname)
library(sp)
library(sf)
data(meuse)
coordinates(meuse) = ~x+y

因此，我从那里开始有接下来的两组A和B点A和B

A <- st_as_sf(meuse) %>% select(geometry) #155 obs
B <- filter(A, row_number()<= 100)%>% select(geometry) #100 obs

，我想挽救A中不在B（即剩下的55分）中的A中的点。

Answer 1

看起来您可以在％操作员中使用％或st_difference函数。如果使用st_difference，则可能需要ST_COMBINE单独设置，如果这些点为sf，则不是sfc sfc代码>。

这些方法可能会因您使用的数据类型而有所不同。它们只是点（SFC），还是与数据（SF）相关的点？

*Edit for your reproducible example:

library(dplyr)
library(ggplot2)
library(sp)
library(sf)

data(meuse)
coordinates(meuse) = ~x+y

A <- st_as_sf(meuse) %>% select(geometry) #155 obs
B <- filter(A, row_number()<= 100)%>% select(geometry) #100 obs

diff_meuse <- st_difference(st_combine(A), st_combine(B)) %>% st_cast('POINT')

diff_meuse
#> Geometry set for 55 features 
#> Geometry type: POINT
#> Dimension:     XY
#> Bounding box:  xmin: 178786 ymin: 329714 xmax: 180923 ymax: 333116
#> CRS:           NA
#> First 5 geometries:
#> POINT (178786 329822)
#> POINT (178875 330311)
#> POINT (179024 329733)
#> POINT (179030 330082)
#> POINT (179034 330561)

ggplot() +
  geom_sf(data = A, size = 4, alpha = .3) + 
  geom_sf(data = diff_meuse, color = 'red')

^{Created on 2022-03-31 by the reprex package (v2.0.1)}

``` r
library(sf)
library(ggplot2)
set.seed(10)

# using nc data incl w/sf
nc <- st_read(system.file("shape/nc.shp", package="sf"))
#> Reading layer `nc' from data source  
#>   using driver `ESRI Shapefile'
#> Simple feature collection with 100 features and 14 fields
#> Geometry type: MULTIPOLYGON
#> Dimension:     XY
#> Bounding box:  xmin: -84.32385 ymin: 33.88199 xmax: -75.45698 ymax: 36.58965
#> Geodetic CRS:  NAD27

a_b <- st_sample(nc, size = 20)

a <- a_b[1:20] 
b <- a_b[1:5]  

diff <- a[!a%in%b]
head(diff)
#> Geometry set for 6 features 
#> Geometry type: POINT
#> Dimension:     XY
#> Bounding box:  xmin: -81.98142 ymin: 34.15653 xmax: -78.5457 ymax: 36.41477
#> Geodetic CRS:  NAD27
#> First 5 geometries:
#> POINT (-78.5457 34.15653)
#> POINT (-79.28979 36.045)
#> POINT (-79.03986 36.41477)
#> POINT (-80.52165 35.15185)
#> POINT (-81.98142 35.45012)

## or using st_difference
diff_st <- st_difference(st_combine(a), st_combine(b))
head(st_cast(diff_st, 'POINT'))
#> Geometry set for 6 features 
#> Geometry type: POINT
#> Dimension:     XY
#> Bounding box:  xmin: -82.19945 ymin: 35.11489 xmax: -80.52165 ymax: 35.93451
#> Geodetic CRS:  NAD27
#> First 5 geometries:
#> POINT (-82.19945 35.93451)
#> POINT (-80.57638 35.60809)
#> POINT (-80.72523 35.14048)
#> POINT (-80.78783 35.11489)
#> POINT (-80.52165 35.15185)

ggplot() + 
  geom_sf(data = diff, color = "black", alpha = .2, size = 6) +
  geom_sf(data = diff_st, color = "black", alpha = .8, size = 4) + 
  geom_sf(data = a, color = 'red', size = 1)

^{Created on 2022-03-31 by the reprex package (v2.0.1)< /sup>}

Answer 2

使用 st_difference 完全没问题，但是在某些情况下，使用空间操作可能会很昂贵（时间和内存）。

因此，对于此示例，您可以考虑将数据视为 data.frame，而不是空间对象，并使用“反连接”

library(data.table)
library(sfheaders)

## Convert to data.frame/data.table
dfA <- sfheaders::sf_to_df(A, fill = T)
dfB <- sfheaders::sf_to_df(B, fill = T)

setDT(dfA)
setDT(dfB)

## do an 'anti-join' to find those in A but not in B 
dfA[ !dfB, on = .(x, y) ]

## I prefer using data.table, but the same anti-join exists in dplyr
dplyr::anti_join(dfA, dfB, by = c("x","y"))


library(ggplot2)

ggplot() +
  geom_sf(data = A, size = 4, alpha = 0.3) +
  geom_point(data = dfA[ !dfB, on = .(x, y) ], aes(x = x, y = y), colour = 'red')