CODEWITHSUNDEEP
jsonrlist
CptNemo
CptNemo

Reputation: 6755

R: Convert list of lists with varying number of items to dataframe

List-to-dataframe have been extensively discussed here but I couldn't find a solution for quite a simple problem: a list of lists with varying number of items.

This is the list

require(json)
hdi <- fromJSON(file="http://data.undp.org/resource/y8j2-3vi9.json")

with the first list of the list being

hdi[1]
[[1]]
[[1]]$`_2011_hdi_value`
[1] "0.887"

[[1]]$`_1990_hdi_value`
[1] "0.798"

[[1]]$`_2000_2013_average_annual_hdi_growth`
[1] "0.37"

[[1]]$`_1980_hdi_value`
[1] "0.757"

[[1]]$`_2010_hdi_value`
[1] "0.885"

[[1]]$`_2008_hdi_value`
[1] "0.879"

[[1]]$`_1990_2000_average_annual_hdi_growth`
[1] "0.62"

[[1]]$`_2012_hdi_value`
[1] "0.889"

[[1]]$`_2013_hdi_value`
[1] "0.890"

[[1]]$`_2005_hdi_value`
[1] "0.870"

[[1]]$`_2000_hdi_value`
[1] "0.849"

[[1]]$country
[1] "Very high human development"

[[1]]$`_1980_1990_average_annual_hdi_growth`
[1] "0.52"

yet the number of items across lists varies

summary(as.numeric(summary(hdi)[,"Length"]))
#   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
#   5.00   16.00   16.00   14.93   16.00   16.00

I want to build a dataframe with 16 columns with NA indicating the missing value for that list.

Upvotes: 1

Views: 597

Answers (2)

Khashaa
Khashaa

Reputation: 7373

Easiest is

data <- jsonlite::fromJSON("http://data.undp.org/resource/y8j2-3vi9.json")

Or, using rlist

library(rlist)
data <- list.stack(hdi, fill=TRUE)

Upvotes: 4

nicola
nicola

Reputation: 24480

You can try this:

 require(RJSONIO)
 hdi <- fromJSON("http://data.undp.org/resource/y8j2-3vi9.json")
 #get the unique values of the column names
 columnnames<-unique(unlist(lapply(hdi,names)))
 #subset each element of the list with and rbind them together
 res<-do.call(rbind,lapply(hdi,function(x) x[columnnames]))
 colnames(res)<-columnnames

From there, you can coerce the numeric column and coerce the object to a data.frame. I suggest you to use the RJSONIO library to handle conversion from/to json as it seems to be more reliable in my experience.

Upvotes: 1

Related Questions