通过绑定数据帧为R进行更快的for循环

Question

I am trying to bind dataframes that comes from JSON data

I tried using rbind.fill and for loop which works for small data, but it takes too long for data more than 100k. Especially, I would like to know if there is any way to vectorize to make it faster rather than making an empty dataframe.

big[1,1] shows a list of json string looks like below

"[{\"latitude\":3750772,\"longitude\":12714673},
{\"latitude\":3750957,\"longitude\":12714793},
{\"latitude\":3751111,\"longitude\":12714954},
{\"latitude\":3751215,\"longitude\":12715155},
{\"latitude\":3751174,\"longitude\":12715295},
{\"latitude\":3751153,\"longitude\":12715174}]"

fromJSONbig[1,1] shows a 6 x 2 dataframe.

library(jsonlite)
library(plyr)

big=fromJSON('RT_data_this_should_be_used_for_rt_analysis.json')
big[1,1]
fromJSON(big[1,1]) #It is a 6 x 2 dataframe

row=nrow(big) #Number of row which also means number of 'rt's

result=data.frame(latitude=integer(), longitude=integer()) #Make an empty dataframe which will store values
for (i in 1:row){
  result=rbind.fill(result,fromJSON(big[i,1])) #Bind the dataframes
}
result[,1]=result[,1]/100000
result[,2]=result[,2]/100000 #Adjust longitude and latitude
result #It would be 6*row x 2 dataframe

Answer 1

This is untested, but maybe something similar would work:

result_list <- lapply(big[, 1], "fromJSON")
result <- do.call("rbind.fill", result_list)

Answer 2

Probably not the most elegant answer, but you could read the arrays into a list and then use the reduce function to bind all the rows together.

resultlist <- vector(list, row)

for (i in 1:row){
  resultlist[[i]]= fromJSON(big[i,1]))
}

result <- reduce(resultlist, rbind.fill)

I expect this should be way faster since the dataframe is not being enlarged in every loop.