There are bunch of files in a directory that has json formatted entries in each line. The size of the files varies from 5k to 200MB. I have this code to go though each file, parse the data I am looking for in the json and finally form a data frame. This script is taking a very long time to finish, in fact it never finishes.
Is there any way to speed it up so that I can read the files faster?
Code:
library(jsonlite)
library(data.table)
setwd("C:/Files/")
#data <- lapply(readLines("test.txt"), fromJSON)
df<-data.frame(Timestamp=factor(),Source=factor(),Host=factor(),Status=factor())
filenames <- list.files("Json_files", pattern="*.txt", full.names=TRUE)
for(i in filenames){
print(i)
data <- lapply(readLines(i), fromJSON)
myDf <- do.call("rbind", lapply(data, function(d) {
data.frame(TimeStamp = d$payloadData$timestamp,
Source = d$payloadData$source,
Host = d$payloadData$host,
Status = d$payloadData$status)}))
df<-rbind(df,myDf)
}
This is a sample entry but there are thousands of entries like this in the file:
{"senderDateTimeStamp":"2016/04/08 10:53:18","senderHost":null,"senderAppcode":"app","senderUsecase":"appinternalstats_prod","destinationTopic":"app_appinternalstats_realtimedata_topic","correlatedRecord":false,"needCorrelationCacheCleanup":false,"needCorrelation":false,"correlationAttributes":null,"correlationRecordCount":0,"correlateTimeWindowInMills":0,"lastCorrelationRecord":false,"realtimeESStorage":true,"receiverDateTimeStamp":1460127623591,"payloadData":{"timestamp":"2016-04-08T10:53:18.169","status":"get","source":"STREAM","fund":"JVV","client":"","region":"","evetid":"","osareqid":"","basis":"","pricingdate":"","content":"","msgname":"","recipient":"","objid":"","idlreqno":"","host":"WEB01","servermember":"test"},"payloadDataText":"","key":"app:appinternalstats_prod","destinationTopicName":"app_appinternalstats_realtimedata_topic","hdfsPath":"app/appinternalstats_prod","esindex":"app","estype":"appinternalstats_prod","useCase":"appinternalstats_prod","appCode":"app"}
{"senderDateTimeStamp":"2016/04/08 10:54:18","senderHost":null,"senderAppcode":"app","senderUsecase":"appinternalstats_prod","destinationTopic":"app_appinternalstats_realtimedata_topic","correlatedRecord":false,"needCorrelationCacheCleanup":false,"needCorrelation":false,"correlationAttributes":null,"correlationRecordCount":0,"correlateTimeWindowInMills":0,"lastCorrelationRecord":false,"realtimeESStorage":true,"receiverDateTimeStamp":1460127623591,"payloadData":{"timestamp":"2016-04-08T10:53:18.169","status":"get","source":"STREAM","fund":"JVV","client":"","region":"","evetid":"","osareqid":"","basis":"","pricingdate":"","content":"","msgname":"","recipient":"","objid":"","idlreqno":"","host":"WEB02","servermember":""},"payloadDataText":"","key":"app:appinternalstats_prod","destinationTopicName":"app_appinternalstats_realtimedata_topic","hdfsPath":"app/appinternalstats_prod","esindex":"app","estype":"appinternalstats_prod","useCase":"appinternalstats_prod","appCode":"app"}
{"senderDateTimeStamp":"2016/04/08 10:55:18","senderHost":null,"senderAppcode":"app","senderUsecase":"appinternalstats_prod","destinationTopic":"app_appinternalstats_realtimedata_topic","correlatedRecord":false,"needCorrelationCacheCleanup":false,"needCorrelation":false,"correlationAttributes":null,"correlationRecordCount":0,"correlateTimeWindowInMills":0,"lastCorrelationRecord":false,"realtimeESStorage":true,"receiverDateTimeStamp":1460127623591,"payloadData":{"timestamp":"2016-04-08T10:53:18.169","status":"get","source":"STREAM","fund":"JVV","client":"","region":"","evetid":"","osareqid":"","basis":"","pricingdate":"","content":"","msgname":"","recipient":"","objid":"","idlreqno":"","host":"WEB02","servermember":""},"payloadDataText":"","key":"app:appinternalstats_prod","destinationTopicName":"app_appinternalstats_realtimedata_topic","hdfsPath":"app/appinternalstats_prod","esindex":"app","estype":"appinternalstats_prod","useCase":"appinternalstats_prod","appCode":"app"}
With your example data in "c:/tmp.txt":
> df <- jsonlite::fromJSON(paste0("[",paste0(readLines("c:/tmp.txt"),collapse=","),"]"))$payloadData[c("timestamp","source","host","status")]
> df
timestamp source host status
1 2016-04-08T10:53:18.169 STREAM WEB01 get
2 2016-04-08T10:53:18.169 STREAM WEB02 get
3 2016-04-08T10:53:18.169 STREAM WEB02 get
So to adapt your code to get a list of dataframes:
dflist <- lapply(filenames, function(i) {
jsonlite::fromJSON(
paste0("[",
paste0(readLines(i),collapse=","),
"]")
)$payloadData[c("timestamp","source","host","status")]
})
The idea is to transform your lines (from readLines) into a big json array and then create the dataframe by parsing it as json.
As lmo already showcased, using lapply on your filenmaes list procide you with a list of dataframes, if you really want only one dataframe at end you can load the data.table packages and then use rbindlist on dflist to get only one dataframe.
Or if you're short in memory this thread may help you.
One speed up is to replace your for loop with lapply Then drop the final rbind. the speed up here would be that R would not have to repeatedly copy an increasingly large file, df over your "bunch" of files. The result would be stored in a convenient list that you could either use as is or convert to a data.frame in one go:
# create processing function
getData <- function(i) {
print(i)
data <- lapply(readLines(i), fromJSON)
myDf <- do.call("rbind", lapply(data, function(d) {
data.frame(TimeStamp = d$payloadData$timestamp,
Source = d$payloadData$source,
Host = d$payloadData$host,
Status = d$payloadData$status)}))
}
# lapply over files
myDataList <- lapply(filenames, getData)
I have some JSON that looks like this:
"total_rows":141,"offset":0,"rows":[
{"id":"1","key":"a","value":{"SP$Sale_Price":"240000","CONTRACTDATE$Contract_Date":"2006-10-26T05:00:00"}},
{"id":"2","key":"b","value":{"SP$Sale_Price":"2000000","CONTRACTDATE$Contract_Date":"2006-08-22T05:00:00"}},
{"id":"3","key":"c","value":{"SP$Sale_Price":"780000","CONTRACTDATE$Contract_Date":"2007-01-18T06:00:00"}},
...
In R, what would be the easiest way to produce a scatter-plot of SP$Sale_Price versus CONTRACTDATE$Contract_Date?
I got this far:
install.packages("rjson")
library("rjson")
json_file <- "http://localhost:5984/testdb/_design/sold/_view/sold?limit=100"
json_data <- fromJSON(file=json_file)
install.packages("plyr")
library(plyr)
asFrame <- do.call("rbind.fill", lapply(json_data, as.data.frame))
but now I'm stuck...
> plot(CONTRACTDATE$Contract_Date, SP$Sale_Price)
Error in plot(CONTRACTDATE$Contract_Date, SP$Sale_Price) :
object 'CONTRACTDATE' not found
How to make this work?
Suppose you have the following JSON-file:
txt <- '{"total_rows":141,"offset":0,"rows":[
{"id":"1","key":"a","value":{"SP$Sale_Price":"240000","CONTRACTDATE$Contract_Date":"2006-10-26T05:00:00"}},
{"id":"2","key":"b","value":{"SP$Sale_Price":"2000000","CONTRACTDATE$Contract_Date":"2006-08-22T05:00:00"}},
{"id":"3","key":"c","value":{"SP$Sale_Price":"780000","CONTRACTDATE$Contract_Date":"2007-01-18T06:00:00"}}]}'
Then you can read it as follows with the jsonlite package:
library(jsonlite)
json_data <- fromJSON(txt, flatten = TRUE)
# get the needed dataframe
dat <- json_data$rows
# set convenient names for the columns
# this step is optional, it just gives you nicer columnnames
names(dat) <- c("id","key","sale_price","contract_date")
# convert the 'contract_date' column to a datetime format
dat$contract_date <- strptime(dat$contract_date, format="%Y-%m-%dT%H:%M:%S", tz="GMT")
Now you can plot:
plot(dat$contract_date, dat$sale_price)
Which gives:
If you choose not to flatten the JSON, you can do:
json_data <- fromJSON(txt)
dat <- json_data$rows$value
sp <- strtoi(dat$`SP$Sale_Price`)
cd <- strptime(dat$`CONTRACTDATE$Contract_Date`, format="%Y-%m-%dT%H:%M:%S", tz="GMT")
plot(cd,sp)
Which gives the same plot:
I found a way that doesn't discard the field names:
install.packages("jsonlite")
install.packages("curl")
json <- fromJSON(json_file)
r <- json$rows
At this point r looks like this:
> class(r)
[1] "data.frame"
> colnames(r)
[1] "id" "key" "value"
After some more Googling and trial-and-error I landed on this:
f <- r$value
sp <- strtoi(f[["SP$Sale_Price"]])
cd <- strptime(f[["CONTRACTDATE$Contract_Date"]], format="%Y-%m-%dT%H:%M:%S", tz="GMT")
plot(cd,sp)
And the result on my full data-set...
I have a csv file and one of the column is in json format.
that particular column in json format looks like this:
{"title":" ","body":" ","url":"thedailygreen print this healthy eating eat safe Dirty Dozen Foods page all"}
I have read this file using read.csv in R. Now, how to I create a new data frame from this column which should have field names as title, body and url.
You can use package RJSONIO to parse the column values, e.g. :
library(RJSONIO)
# create an example data.frame with a json column
cell1 <- '{"title":"A","body":"X","url":"http://url1.x"}'
cell2 <- '{"title":"B","body":"Y","url":"http://url2.y"}'
cell3 <- '{"title":"C","body":"Z","url":"http://url3.z"}'
df <- data.frame(jsoncol = c(cell1,cell2,cell3),stringsAsFactors=F)
# parse json and create a data.frame
res <- do.call(rbind.data.frame,
lapply(df$jsoncol, FUN=function(x){ as.list(fromJSON(x))}))
> res
title body url
A X http://url1.x
B Y http://url2.y
C Z http://url3.z
N.B. :
the code above assumes all the cells contains title, body and url only. If there can be other properties in the json cells, use this code instead :
vals <- lapply(df$jsoncol,fromJSON)
res <- do.call(rbind, lapply(vals,FUN=function(v){ data.frame(title=v['title'],
body =v['body'],
url =v['url']) }))
EDIT (as per comment):
I've read the file using the following code :
df <- read.table(file="c:\\sample.tsv",
header=T, sep="\t", colClasses="character")
then parsed using this code :
# define a simple function to turn NULL to NA
naIfnull <- function(x){if(!is.null(x)) x else NA}
vals <- lapply(df$boilerplate,fromJSON)
res <- do.call(rbind,
lapply(vals,FUN=function(v){ v <- as.list(v)
data.frame(title=naIfnull(v$title),
body =naIfnull(v$body),
url =naIfnull(v$url)) }))
I'm trying to download NBA player information from Numberfire and then put that information into a data frame. However I seem to be running into a few issues
The following snippet downloads the information just fine
require(RCurl)
require(stringr)
require(rjson)
#download data from numberfire
nf <- "https://www.numberfire.com/nba/fantasy/fantasy-basketball-projections"
html <- getURL(nf)
Then there is what I assume to be a JSON data structure
#extract json variable (?)
pat <- "NF_DATA.*}}}"
jsn <- str_extract(html, pat)
jsn <- str_split(jsn, "NF_DATA = ")
parse <- newJSONParser()
parse$addData(jsn)
It seems to add data OK as it doesn't throw any errors, but if there is data in that object I can't tell or seem to get it out!
I'd paste in the jsn variable but it's way over the character limit. Any hints as to where I'm going wrong would be much appreciated
Adding the final line gets a nice list format that you can transform to a data.frame
require(RCurl); require(stringr); require(rjson)
#download data from numberfire
nf <- "https://www.numberfire.com/nba/fantasy/fantasy-basketball-projections"
html <- getURL(nf)
#extract json variable (?)
pat <- "NF_DATA.*}}}"
jsn <- str_extract(html, pat)
jsn <- str_split(jsn, "NF_DATA = ")
fromJSON(jsn[[1]][[2]])
I am trying to use R's rjson library to upload a 200MB JSON file into R , but I got the Cannot fit vector over 1KB error.
Here's the code I used to load the JSON file into R:
UnpackJSON <- function(filePath)
{
con <- file(filePath, "r")
input <- readLines(con, -1L)
# jsonData <- fromJSON(paste(input, collapse=""))
jsonData <- sapply(input, fromJSON)
close(con)
df <- data.frame(jsonData)
temp <- rownames(df)
df <- as.data.frame(t(df))
colnames(df) <- temp
rownames(df) <- NULL
return(df)
}
Is there a way to optimize this code or another way to load such a large file into R? I appreciate any input.
Why would you process the JSON data line per line using sapply? Couldn't you just say fromJSON(input)? I suspect a problem could be that the garbage collector waits for the entire sapply loop to finish before deleting any copies that are made in the analysis process (see also the comments here). So when vectorization does not work (which is greatly preferred), you could try and change the sapply loop to a for loop and see if that helps.