Removing a string using regular expression with stringr

Question

I am trying to get rid of the term union in the following json using str_replace_all function. I tried few regular expressions but they didn't work. Any suggestions on how to get rid of the full string of "union". Here is the code.

 library(jsonlite)

 json<- '{"open_issues":0,"union":{"avatar_url":"https://secure.gravatar.com/avatar/ed73cf7f24c091318e507d167c7b3456?d=https://a248.e.akamai.net/assets.mashups.com%2Fimages%2Fgravatars%2Fgravatar-orgs.png","gravatar_id":"ed73cf7f24c091318e507d167c7b3456","id":1556245,"login":"Toolpark","url":"https://api.mashups.com/users/Toolpark"},"owner":{"avatar_url":"https://secure.gravatar.com/avatar/ed73cf7f24c091318e507d167c7b3456?d=https://a248.e.akamai.net/assets.mashups.com%2Fimages%2Fgravatars%2Fgravatar-orgs.png","gravatar_id":"ed73cf7f24c091318e507d167c7b3456","id":1556245,"login":"Toolpark","url":"https://api.mashups.com/users/Toolpark"},"private":false,"pushed_at":"2012-07-12T11:43:31Z","size":964,"ssh_url":"vit@mashups.com:Toolpark/irma.vit","svn_url":"https://mashups.com/Toolpark/irma","updated_at":"2012-07-12T11:43:31Z","url":"https://api.mashups.com/repos/Toolpark/irma","watchers":2}'

str_replace_all(json,"union[^}]*}\s*, "")
json <- fromJSON(json)
Error: '\s' is an unrecognized escape in character string starting ""union[^}]*}\s"

I am trying to get rid of this string exactly including the coma in the end:

"union":{"avatar_url":"https://secure.gravatar.com/avatar/ed73cf7f24c091318e507d167c7b3456?d=https://a248.e.akamai.net/assets.mashups.com%2Fimages%2Fgravatars%2Fgravatar-orgs.png","gravatar_id":"ed73cf7f24c091318e507d167c7b3456","id":1556245,"login":"Toolpark","url":"https://api.mashups.com/users/Toolpark"},

Any suggestions?

it looks like you are missing some quotation marks on the str_replace_all line — Smern, Jun 22 '15 at 18:45
Try `str_replace_all(json, '"union"[^}]*\\s*,', "")`. Keep in mind that this may not work with stripping newlines between `}` and `,` (not that that's generally a concern). Also keep in mind that if `}` is in the `union` string it won't work like you expect either. It would make more sense to parse the JSON and then delete the `union` property. — Explosion Pills, Jun 22 '15 at 18:48
Maybe I missed something but, why you just don't `cat(gsub("union", "", json))` ? — SabDeM, Jun 22 '15 at 18:50
Try this: sub('.*(?<=Toolpark\\/)(irma.*)','\\1',string,perl=T) — Shenglin Chen, Jun 22 '15 at 18:55
I need a general regular expression which can get rid of more strings of this nature. — user3570187, Jun 22 '15 at 18:57
I am using `code` file <- "json.json" library('tm') json <- readLines(json). And i issued the command you suggested and i am still getting the union column. — user3570187, Jun 22 '15 at 19:05
Your example works for me if I use `json <- fromJSON(gsub("union", "", json))`. — RHertel, Jun 22 '15 at 19:26

score 2 · Accepted Answer · answered Jun 22 '15 at 21:06

2

If I am guessing correct, you intend to delete the "union" element from the json line. So, why don't you change the line into json using fromJSON function and then try deleting the union element. The following code yields you a list without union

json1 <- ("your json line") 
my_list<- fromJSON(json1)   #convert into json format which gives you a list of elements
my_list["union"]<- NULL   #delete the union element
print(my_list)

answered Jun 22 '15 at 21:06

Shiva

789
6
15

i have a huge json file which is difficult to be copied into the .R file and i am seeking some alternative solution. Note that I don't have a valid json file and i need to make it to valid json and do the operations! Thanks so much! – user3570187 Jun 23 '15 at 00:08

score 1 · Answer 2 · answered Jun 23 '15 at 03:17

1

You have been using your regex with bad syntax.. Use the following:

str_replace_all(json, "\"union[^}]*}\\s*,", "")

See DEMO

answered Jun 23 '15 at 03:17

karthik manchala

13,492
1
31
55

1

Thank you so much! I am getting this error: Error in stri_replace_all_regex(string, pattern, replacement, vectorize_all = vec, : Syntax error in regexp pattern. (U_REGEX_RULE_SYNTAX). Not sure if it is a r package issue. – user3570187 Jun 23 '15 at 13:48

Removing a string using regular expression with stringr

2 Answers2