将嵌套列表转换为具有不同列长度的 data.frame
Convert Nested List into data.frame with different column length
我试图将下面的嵌套列表转换为 data.frame 但没有成功。有一些并发症，主要是位置1的“结果”列与位置2不一致，因为位置2没有结果。
不同位置的项目长度不一致
[[1]]
[[1]]$html_attributions
list()
[[1]]$results
geometry.location.lat geometry.location.lng
1 25.66544 -100.4354
id place_id
1 6ce0a030663144c8e992cbce51eb00479ef7db89 ChIJVy7b7FW9YoYRdaH2I_gOJIk
reference
1 CmRSAAAATdtVfB4Tz1aQ8GhGaw4-nRJ5lZlVNgiOR3ciF4QjmYC56bn6b7omWh1SJEWWqQQEFNXxGZndgEwSgl8sRCOtdF8aXpngUY878Q__yH4in8EMZMCIqSHLARqNgGlV4mKgEhDlvkHLXLiBW4F_KQVT83jIGhS5DJipk6PAnpPDXP2p-4X5NPuG9w
[[1]]$status
[1] "OK"
[[2]]
[[2]]$html_attributions
list()
[[2]]$results
list()
[[2]]$status
[1] "ZERO_RESULTS"
我尝试了以下代码，但它们不起作用。
#1
m1 <- do.call(rbind, lapply(myDataFrames, function(y) do.call(rbind, y)))
relist(m1, skeleton = myDataFrames)
#2
relist(matrix(unlist(myDataFrames), ncol = 4, byrow = T), skeleton = myDataFrames)
#3
library(data.table)
df<-rbindlist(myDataFrames, idcol = "index")
df<-rbindlist(myDataFrames, fill=TRUE)
#4
myDataFrame <- do.call(rbind.data.frame, c(myDataFrames, list(stringsAsFactors = FALSE)))
我想我有足够的原始资料 JSON 可以创建一个可重现的例子：
okjson <- '{"html_attributions":[],"results":[{"geometry":{"location":{"lat":25.66544,"lon":-100.4354},"id":"foo","place_id":"quux"}}],"status":"OK"}'
emptyjson <- '{"html_attributions":[],"results":[],"status":"ZERO_RESULTS"}'
jsons <- list(okjson, emptyjson, okjson)
从这里开始，我将逐步（缓慢地）完成整个过程。为了重现性，我已经包含了很多中间结构，对于冗长，我深表歉意。这可以很容易地组合在一起 and/or 放在 magrittr 管道中。
lists <- lapply(jsons, jsonlite::fromJSON)
str(lists)
# List of 3
# $ :List of 3
# ..$ html_attributions: list()
# ..$ results :'data.frame': 1 obs. of 1 variable:
# .. ..$ geometry:'data.frame': 1 obs. of 3 variables:
# .. .. ..$ location:'data.frame': 1 obs. of 2 variables:
# .. .. .. ..$ lat: num 25.7
# .. .. .. ..$ lon: num -100
# .. .. ..$ id : chr "foo"
# .. .. ..$ place_id: chr "quux"
# ..$ status : chr "OK"
# $ :List of 3
# ..$ html_attributions: list()
# ..$ results : list()
# ..$ status : chr "ZERO_RESULTS"
# $ :List of 3
# ..$ html_attributions: list()
# ..$ results :'data.frame': 1 obs. of 1 variable:
# .. ..$ geometry:'data.frame': 1 obs. of 3 variables:
# .. .. ..$ location:'data.frame': 1 obs. of 2 variables:
# .. .. .. ..$ lat: num 25.7
# .. .. .. ..$ lon: num -100
# .. .. ..$ id : chr "foo"
# .. .. ..$ place_id: chr "quux"
# ..$ status : chr "OK"
goodlists <- Filter(function(a) "results" %in% names(a) && length(a$results) > 0, lists)
goodresults <- lapply(goodlists, `[[`, "results")
str(goodresults)
# List of 2
# $ :'data.frame': 1 obs. of 1 variable:
# ..$ geometry:'data.frame': 1 obs. of 3 variables:
# .. ..$ location:'data.frame': 1 obs. of 2 variables:
# .. .. ..$ lat: num 25.7
# .. .. ..$ lon: num -100
# .. ..$ id : chr "foo"
# .. ..$ place_id: chr "quux"
# $ :'data.frame': 1 obs. of 1 variable:
# ..$ geometry:'data.frame': 1 obs. of 3 variables:
# .. ..$ location:'data.frame': 1 obs. of 2 variables:
# .. .. ..$ lat: num 25.7
# .. .. ..$ lon: num -100
# .. ..$ id : chr "foo"
# .. ..$ place_id: chr "quux"
goodresultsdf <- lapply(goodresults, function(a) jsonlite::flatten(as.data.frame(a)))
str(goodresultsdf)
# List of 2
# $ :'data.frame': 1 obs. of 4 variables:
# ..$ geometry.id : chr "foo"
# ..$ geometry.place_id : chr "quux"
# ..$ geometry.location.lat: num 25.7
# ..$ geometry.location.lon: num -100
# $ :'data.frame': 1 obs. of 4 variables:
# ..$ geometry.id : chr "foo"
# ..$ geometry.place_id : chr "quux"
# ..$ geometry.location.lat: num 25.7
# ..$ geometry.location.lon: num -100
我们现在有 list 个 data.frame 个，是个好地方。
do.call(rbind.data.frame, c(goodresultsdf, stringsAsFactors = FALSE))
# geometry.id geometry.place_id geometry.location.lat geometry.location.lon
# 1 foo quux 25.66544 -100.4354
# 2 foo quux 25.66544 -100.4354
我试图将下面的嵌套列表转换为 data.frame 但没有成功。有一些并发症，主要是位置1的“结果”列与位置2不一致，因为位置2没有结果。
不同位置的项目长度不一致
[[1]]
[[1]]$html_attributions
list()
[[1]]$results
geometry.location.lat geometry.location.lng
1 25.66544 -100.4354
id place_id
1 6ce0a030663144c8e992cbce51eb00479ef7db89 ChIJVy7b7FW9YoYRdaH2I_gOJIk
reference
1 CmRSAAAATdtVfB4Tz1aQ8GhGaw4-nRJ5lZlVNgiOR3ciF4QjmYC56bn6b7omWh1SJEWWqQQEFNXxGZndgEwSgl8sRCOtdF8aXpngUY878Q__yH4in8EMZMCIqSHLARqNgGlV4mKgEhDlvkHLXLiBW4F_KQVT83jIGhS5DJipk6PAnpPDXP2p-4X5NPuG9w
[[1]]$status
[1] "OK"
[[2]]
[[2]]$html_attributions
list()
[[2]]$results
list()
[[2]]$status
[1] "ZERO_RESULTS"
我尝试了以下代码，但它们不起作用。
#1
m1 <- do.call(rbind, lapply(myDataFrames, function(y) do.call(rbind, y)))
relist(m1, skeleton = myDataFrames)
#2
relist(matrix(unlist(myDataFrames), ncol = 4, byrow = T), skeleton = myDataFrames)
#3
library(data.table)
df<-rbindlist(myDataFrames, idcol = "index")
df<-rbindlist(myDataFrames, fill=TRUE)
#4
myDataFrame <- do.call(rbind.data.frame, c(myDataFrames, list(stringsAsFactors = FALSE)))
我想我有足够的原始资料 JSON 可以创建一个可重现的例子：
okjson <- '{"html_attributions":[],"results":[{"geometry":{"location":{"lat":25.66544,"lon":-100.4354},"id":"foo","place_id":"quux"}}],"status":"OK"}'
emptyjson <- '{"html_attributions":[],"results":[],"status":"ZERO_RESULTS"}'
jsons <- list(okjson, emptyjson, okjson)
从这里开始，我将逐步（缓慢地）完成整个过程。为了重现性，我已经包含了很多中间结构，对于冗长，我深表歉意。这可以很容易地组合在一起 and/or 放在 magrittr 管道中。
lists <- lapply(jsons, jsonlite::fromJSON)
str(lists)
# List of 3
# $ :List of 3
# ..$ html_attributions: list()
# ..$ results :'data.frame': 1 obs. of 1 variable:
# .. ..$ geometry:'data.frame': 1 obs. of 3 variables:
# .. .. ..$ location:'data.frame': 1 obs. of 2 variables:
# .. .. .. ..$ lat: num 25.7
# .. .. .. ..$ lon: num -100
# .. .. ..$ id : chr "foo"
# .. .. ..$ place_id: chr "quux"
# ..$ status : chr "OK"
# $ :List of 3
# ..$ html_attributions: list()
# ..$ results : list()
# ..$ status : chr "ZERO_RESULTS"
# $ :List of 3
# ..$ html_attributions: list()
# ..$ results :'data.frame': 1 obs. of 1 variable:
# .. ..$ geometry:'data.frame': 1 obs. of 3 variables:
# .. .. ..$ location:'data.frame': 1 obs. of 2 variables:
# .. .. .. ..$ lat: num 25.7
# .. .. .. ..$ lon: num -100
# .. .. ..$ id : chr "foo"
# .. .. ..$ place_id: chr "quux"
# ..$ status : chr "OK"
goodlists <- Filter(function(a) "results" %in% names(a) && length(a$results) > 0, lists)
goodresults <- lapply(goodlists, `[[`, "results")
str(goodresults)
# List of 2
# $ :'data.frame': 1 obs. of 1 variable:
# ..$ geometry:'data.frame': 1 obs. of 3 variables:
# .. ..$ location:'data.frame': 1 obs. of 2 variables:
# .. .. ..$ lat: num 25.7
# .. .. ..$ lon: num -100
# .. ..$ id : chr "foo"
# .. ..$ place_id: chr "quux"
# $ :'data.frame': 1 obs. of 1 variable:
# ..$ geometry:'data.frame': 1 obs. of 3 variables:
# .. ..$ location:'data.frame': 1 obs. of 2 variables:
# .. .. ..$ lat: num 25.7
# .. .. ..$ lon: num -100
# .. ..$ id : chr "foo"
# .. ..$ place_id: chr "quux"
goodresultsdf <- lapply(goodresults, function(a) jsonlite::flatten(as.data.frame(a)))
str(goodresultsdf)
# List of 2
# $ :'data.frame': 1 obs. of 4 variables:
# ..$ geometry.id : chr "foo"
# ..$ geometry.place_id : chr "quux"
# ..$ geometry.location.lat: num 25.7
# ..$ geometry.location.lon: num -100
# $ :'data.frame': 1 obs. of 4 variables:
# ..$ geometry.id : chr "foo"
# ..$ geometry.place_id : chr "quux"
# ..$ geometry.location.lat: num 25.7
# ..$ geometry.location.lon: num -100
我们现在有 list 个 data.frame 个，是个好地方。
do.call(rbind.data.frame, c(goodresultsdf, stringsAsFactors = FALSE))
# geometry.id geometry.place_id geometry.location.lat geometry.location.lon
# 1 foo quux 25.66544 -100.4354
# 2 foo quux 25.66544 -100.4354