r - 将几何图形的字符向量转换为 sfc_LINESTRING

Question

我有一个数据表如下，我需要将其转换为sf对象:

library(data.table)
DT <- data.table(
    ID = c("A", "B", "C", "D", "E"),
    Value = 11:15
)

每个ID的几何信息是一个字符向量，具有不同数量的坐标组成线串。 ID“B”有4个坐标； ID“C”有 5。

geo <- c("c(-112.116492521272, -112.116492811159, -112.116492812107, -112.116491781569, -112.116482854256, -112.116482819195, -112.116476331207, -112.116476325101, -112.11647589954, 33.3777109072744, 33.377733456163, 33.377733512504, 33.377817189599, 33.3785425053239, 33.3785454379367, 33.3790725760563, 33.3790731291841, 33.3791076444333)", 
         "c(-112.282916223332, -112.282955145531, -112.282977080374, -112.282986066594, 33.499285198973, 33.4994146786288, 33.4995335119373, 33.4998030580162)", 
         "c(-112.281058674957, -112.281058522318, -112.281057917087, -112.281057356648, -112.281055594103, -112.281047371356, -112.281048086137, -112.28104821173, 33.4937123457776, 33.4937301348982, 33.4938008007847, 33.4938659107566, 33.4940708243904, 33.4950232493953, 33.4951159682343, 33.4951322463168)", 
         "c(-112.282978024041, -112.282977000088, -112.282975472281, -112.282975387447, -112.282974470679, -112.282974464144, -112.282974284899, -112.28297410899, -112.282974107453, 33.5011764123633, 33.5013710145493, 33.5016617311961, 33.501678000948, 33.5018530730796, 33.5018546369058, 33.5018887965849, 33.5019223852857, 33.5019226044706)", 
         "c(-112.282986066594, -112.282985540911, -112.282984156895, -112.282983004093, -112.282982201845, 33.4998030580162, 33.4998965204233, 33.5001425170464, 33.5003478058912, 33.5004906801949)"
)

将几何信息添加到DT:

DT$geometry <- geo

现在，我需要将 DT 转换为几何形状指定为 sfc_LINESTRING 的 sf 对象。我尝试使用 st_cast 首先将基于字符的几何变量转换为线字符串，但它产生了错误。

DT_sf <- st_cast(DT$geometry, "LINESTRING")

Error in UseMethod("st_cast") : 
  no applicable method for 'st_cast' applied to an object of class "character"

需要对近 20,000 行进行此转换。因此，我正在寻找一种计算有效的方法来实现所需的结果。

Answer 1

您可以使用eval “评估以字符串形式给出的表达式”

(如果您的字符串没有经过净化，这可能会很危险(想想 SQL 注入(inject)))

因此，使用您的 geo 对象，您将得到

lst <- lapply(geo, function(x) { eval(parse(text = x)) })

str( lst )
List of 5
# $ : num [1:18] -112 -112 -112 -112 -112 ...
# $ : num [1:8] -112.3 -112.3 -112.3 -112.3 33.5 ...
# $ : num [1:16] -112 -112 -112 -112 -112 ...
# $ : num [1:18] -112 -112 -112 -112 -112 ...
# $ : num [1:10] -112 -112 -112 -112 -112 ...

由于我们每次评估 geo 中的每个向量(在 lapply 内)，因此我们也可以同时将其设为 sfg 对象

lst <- lapply(geo, function(x) {
  v <- eval(parse(text = x))
  m <- matrix(v, ncol = 2)
  sf::st_linestring(m)
})

然后所需要做的就是添加正确的类属性

DT$geo <- lst

DT$geo <- sf::st_as_sfc( DT$geo )
DT <- sf::st_as_sf( DT )
sf::st_crs( DT ) <- 4326

DT
# Simple feature collection with 5 features and 2 fields
# Geometry type: LINESTRING
# Dimension:     XY
# Bounding box:  xmin: -112.283 ymin: 33.37771 xmax: -112.1165 ymax: 33.50192
# Geodetic CRS:  WGS 84
# ID Value                            geo
# 1  A    11 LINESTRING (-112.1165 33.37...
# 2  B    12 LINESTRING (-112.2829 33.49...
# 3  C    13 LINESTRING (-112.2811 33.49...
# 4  D    14 LINESTRING (-112.283 33.501...
# 5  E    15 LINESTRING (-112.283 33.499...