最快的方法来读取/处理data.table中的“类似于JSON的”列？

Posted 2021-04-27

以下是我正在使用的一些示例数据。 DT_IN包含数据的输入格式，DT_OUT包含我要使用的格式。从DT_IN转到DT_OUT的最佳方法是什么？

我已经尝试过strsplit，但是没有设法按照相应的顺序将拆分分割为rbind。愿意接受解决方案，也许Rcpp可以帮助您吗？

library(data.table)
DT_IN <- data.table(
  user_id = c(1L, 20L, 4L, 6L, 9L), 
  latitude = c(-41.3103218, -40.8307381, -37.3932037, -42.7178726, -45.0156822), 
  longitude = c(174.824554, 172.793106, 175.840637, 170.965454, 168.731186), 
  parameters = c(
    "{""network""=>""Telecom NZ"", ""accuracy""=>28.659999847412, ""internet""=>""4G"", ""location_age""=>1}", 
    "{""location_age""=>716}", "{""location_age""=>851}", "{""accuracy""=>14, ""location_age""=>1}", 
    "{""network""=>""VodafoneNZ"", ""accuracy""=>29, ""internet""=>""3G"", ""location_age""=>31}"
  )
)
> DT_IN
   user_id  latitude longitude                                                                                              parameters
1:       1 -41.31032  174.8246 {""network""=>""Telecom NZ"", ""accuracy""=>28.659999847412, ""internet""=>""4G"", ""location_age""=>1}
2:      20 -40.83074  172.7931                                                                                 {""location_age""=>716}
3:       4 -37.39320  175.8406                                                                                 {""location_age""=>851}
4:       6 -42.71787  170.9655                                                                 {""accuracy""=>14, ""location_age""=>1}
5:       9 -45.01568  168.7312             {""network""=>""VodafoneNZ"", ""accuracy""=>29, ""internet""=>""3G"", ""location_age""=>31}
DT_OUT <- data.table(
  user_id = c(1L, 20L, 4L, 6L, 9L), 
  latitude = c(-41.3103218, -40.8307381, -37.3932037, -42.7178726, -45.0156822), 
  longitude = c(174.824554, 172.793106, 175.840637, 170.965454, 168.731186), 
  network = c('Telecom NZ', NA, NA, NA, 'VodafoneNZ'), 
  accuracy = c(28.659999847412, NA, NA, 14, 29), 
  internet = c('4G', NA, NA, NA, '3G'), 
  location_age = c(1, 716, 851, 1, 31)
)
> DT_OUT
   user_id  latitude longitude    network accuracy internet location_age
1:       1 -41.31032  174.8246 Telecom NZ    28.66       4G            1
2:      20 -40.83074  172.7931       <NA>       NA     <NA>          716
3:       4 -37.39320  175.8406       <NA>       NA     <NA>          851
4:       6 -42.71787  170.9655       <NA>    14.00     <NA>            1
5:       9 -45.01568  168.7312 VodafoneNZ    29.00       3G           31

答案

使用jsonlite包...

# Convert json like strings to json.
DT_IN[, parameters := gsub("""", """, parameters)]
DT_IN[, parameters := gsub("=>", ":", parameters)]

# Stream_in the json and cbind it to existing data.
DT_IN <- cbind(DT_IN, jsonlite::stream_in(textConnection(DT_IN$parameters)))

# Remove `parameters`
DT_IN[, parameters := NULL]

DT_IN
#    user_id  latitude longitude    network accuracy internet location_age
# 1:       1 -41.31032  174.8246 Telecom NZ    28.66       4G            1
# 2:      20 -40.83074  172.7931       <NA>       NA     <NA>          716
# 3:       4 -37.39320  175.8406       <NA>       NA     <NA>          851
# 4:       6 -42.71787  170.9655       <NA>    14.00     <NA>            1
# 5:       9 -45.01568  168.7312 VodafoneNZ    29.00       3G           31