How to fast parse a list of JSON string?

Question

I used rjson and do this JSON <- lapply(data$toParse, fromJSON)

my toParse has 600k rows of simple/short JSON.

However, it is very slow possibly because fromJSON doesn't vectorize operation so I have to use lapply.

I am wondering if there is any better way for parsing a list of JSON?

If nothing has changed for the last few months rjson is the only one json parser available for R. Alternative solution could be parsing json some external tools and preprocessing your data. There is pretty extensive list of available libraries on http://www.json.org/ — zero323, Sep 10 '13 at 18:04

score 1 · Accepted Answer · answered Sep 10 '13 at 19:08

1

What about pasting the text into an array and parsing, rjson::fromJSON(sprintf("[%s]", paste(data$toParse, collapse=",")))?

answered Sep 10 '13 at 19:08

Martin Morgan

I had come up similar idea, but didn't think it would work out much improvement. However, the benchmark shows 30% faster – colinfang Sep 10 '13 at 20:32
so in my case sprintf + paste is faster than lapply + Fun via Rprof – colinfang Sep 10 '13 at 21:41

1 Answers1