Split string of a data frame in new columns - r

Question

I need your help, because I got a data frame with a very difficult format. My data frame
data <- data.frame(information = c("{u'info1': u'mnfd', u'text': u'exampletext'}","{u'info2': u'332', u'text': u'lalala'}","{u'info1': u'', u'text': u'blub'}"))
has the column information (and a few other columns in the real data frame) and looks for example like that
## information
## 1 {u'info1': u'mnfd', u'text': u'exampletext'}
## 2 {u'info2': u'332', u'text': u'lalala'}
## 3 {u'info1': u'', u'text': u'blub'}
The real data frame has a few thousand rows and the strings are much longer. I would like to add columns which should display the information from the strings. So at the end I would like to have a dataframe looking like that (the string "of_" is added before every columnn name)
information of_info1 of_text of_info2
1 {u'info1': u'mnfd', u'text': u'exampletext'} mnfd exampletext <NA>
2 {u'info2': u'332', u'text': u'lalala'} <NA> lalala 322
3 {u'info1': u'', u'text': u'blub'} blub <NA>
Thanks for your help

Answer 1

This is close to a JSON file, so do a bit of formatting to get it right, and then import via the awesome jsonlite package:
library(jsonlite)
fromJSON(paste0("[", paste(gsub("(u|)'",'"',data$information), collapse=",\n"), "]"))
# info1 text info2
#1 mnfd exampletext <NA>
#2 <NA> lalala 332
#3 blub <NA>

Answer 2

Here's a version with dplyr and stringr. It should not be too difficult to translate it into base R if you prefer that.
This will break, however, if there are escaped single quotation marks in the fields.
library(stringr)
library(dplyr)
data <- data$information %>% str_match_all("u'([^']+)': u'([^']*)'") %>%
lapply(function(matches) {
result <- data.frame(as.list(matches[,3]), stringsAsFactors = FALSE)
colnames(result) <- paste0("of_", matches[,2])
result
}) %>% bind_rows() %>% bind_cols(data, .)

Answer 3

Convert it to DCF format and then read it in using read.dcf. No packages are used.
First we remove the junk giving s0 and then split it on comma-space giving s1. Then add an empty terminating line between records giving s2. Finally use read.dcf to read that in and append it to data.
s0 <- gsub("[{}]", "", gsub("u'(.*?)'", "\\1", data$information))
s1 <- strsplit(s0, ", ")
s2 <- unlist(lapply(s1, c, ""))
cbind(data, read.dcf(textConnection(s2)))
giving:
information info1 text info2
1 {u'info1': u'mnfd', u'text': u'exampletext'} mnfd exampletext <NA>
2 {u'info2': u'332', u'text': u'lalala'} <NA> lalala 332
3 {u'info1': u'', u'text': u'blub'} blub <NA>
magrittr
This could also be expressed as a nested magrittr pipeline like this:
library(magrittr)
data %>%
cbind({.$information %>%
gsub("u'(.*?)'", "\\1", .) %>%
gsub("[{}]", "", .) %>%
strsplit(", ") %>%
lapply(c, "") %>%
unlist %>%
textConnection %>%
read.dcf
})