如何使用 R 将数据帧转换为 JSON 数组

Question

在我当前的项目中，我正在尝试从csv 文件中读取数据，并尝试根据R 中csv 文件中的数据创建分层JSON 数组。示例数据如下所示：

添加数据样本数据（为简单起见减少了数据集）：

    Country   Provider   2 G Data   3 G Data    LTE     FP0   anfang0   2G  3G  FP1 anfang1
     ABC        A1          n          n         n      fp0      j      NA  NA  NA  NA
     ABC        A2          NA         NA        NA      NA      NA      j   j  fp1 n
     ABC        A3          n          n         n       fp0     j      NA  NA  NA  NA
     DEF        A7          j          j         j       fp0     n       j   j  fp1 n

对数据的理解：n代表值是no，j代表值是yes，NA代表值缺失。 FP0 和FP1 表示有关同一提供商但在不同区域的信息。单行中有两种类型的数据，即2 G Data, 3 G Data, LTE, FP0, anfang 0 属于一个组，2G, 3G, FP1, anfang 1 属于另一个组。如果所有信息都是n 即no，那么我们必须考虑对应的anfang0 或anfang1 值。

示例输出如下所示（基于上面的解释）：

        {
      "ABC": {
        "fp0":[
          {
            "provider": "A1",
            "anfrage": "j"
          },
          {
            "provider": "A3",
            "anfrage": "j"
          }
        ],
        "fp1": [
          {
            "provider": "A2",
            "2G": "j",
            "3G": "j"
          }
        ]
      },
      "DEF": {
        "fp1": [
          {
            "provider": "A7",
            "2G": "j",
            "3G": "j"
          }  
        ],
        "fp0": [
          {
            "provider": "A7",
            "2G": "j",
            "3G": "j",
            "LTE": "j"
          }  
        ]       
      }  
    }

在上述json 格式中，对于每个Country，应该只有一个json 块，如上所示。到目前为止，我尝试关注this link，但找不到任何可行的解决方案。

for(i in 1:nrow(data)){
   a=c(a,jsonlite::toJSON(list(list('fp0' = 
   list("provider"=data$Provider[i],"2g"=data$`2 G Data`[i],"3g"=data$`3 G 
   Data`[i],"LTE"=data$LTE[i]))), pretty = TRUE))
}
toJSON(a, pretty = TRUE, auto_unbox = TRUE)

如果您需要更清晰的信息，请告诉我。

Answer 1

其中一种方法可能是

library(dplyr)
library(jsonlite)

#data pre-processing (bind different areas' data in row)
df1 <- df[, 1:7] %>%                          #dataframe having data for one area - i.e. fp0
  na.omit() %>%
  `colnames<-`(c("country", "provider", "2G", "3G", "LTE", "fp", "anfang")) %>%
  bind_rows(
    df[, c(1:2, 8:ncol(df))] %>%              #dataframe having data for another area - i.e. fp1
      na.omit() %>%
      `colnames<-`(c("country", "provider", "2G", "3G", "fp", "anfang"))
    )
df1[df1 == 'n'] <- NA                         #convert all "n" to NA as we are not concerened about it in the final output

#convert processed dataframe to a list
dfList <- lapply(split(df1, df1$country), 
                 function(x) split(x[, c("provider", "2G", "3G", "LTE", "anfang")], x$fp))

#final result (convert list to JSON)
json_out <- toJSON(dfList, auto_unbox = T)

给了

> json_out
{"ABC":{"fp0":[{"provider":"A1","anfang":"j"},{"provider":"A3","anfang":"j"}],"fp1":[{"provider":"A2","2G":"j","3G":"j"}]},"DEF":{"fp0":[{"provider":"A7","2G":"j","3G":"j","LTE":"j"}],"fp1":[{"provider":"A7","2G":"j","3G":"j"}]}}

样本数据：

df <- structure(list(Country = c("ABC", "ABC", "ABC", "DEF"), Provider = c("A1", 
"A2", "A3", "A7"), `2 G Data` = c("n", NA, "n", "j"), `3 G Data` = c("n", 
NA, "n", "j"), LTE = c("n", NA, "n", "j"), FP0 = c("fp0", NA, 
"fp0", "fp0"), anfang0 = c("j", NA, "j", "n"), `2G` = c(NA, "j", 
NA, "j"), `3G` = c(NA, "j", NA, "j"), FP1 = c(NA, "fp1", NA, 
"fp1"), anfang1 = c(NA, "n", NA, "n")), .Names = c("Country", 
"Provider", "2 G Data", "3 G Data", "LTE", "FP0", "anfang0", 
"2G", "3G", "FP1", "anfang1"), class = "data.frame", row.names = c(NA, 
-4L))

#  Country Provider 2 G Data 3 G Data  LTE  FP0 anfang0   2G   3G  FP1 anfang1
#1     ABC       A1        n        n    n  fp0       j <NA> <NA> <NA>    <NA>
#2     ABC       A2     <NA>     <NA> <NA> <NA>    <NA>    j    j  fp1       n
#3     ABC       A3        n        n    n  fp0       j <NA> <NA> <NA>    <NA>
#4     DEF       A7        j        j    j  fp0       n    j    j  fp1       n