valid_json() has problems with escaped characters in query() strings

[jeremy-j-ackso]

I'm trying to query for some JSON that's embedded in an ES document as a string. To do that I need to include the quotes internal to that JSON in my regexp search.

A couple very simplified examples of documents I'm searching through for matched text:

{message:'{"product":"toothbrush"}'}
{message:'{"product":"hairspray"}'}
{message:'{"product":"ear swabs"}'}

Here's the query I'm trying to wrap in query():

qry <- '{
  "regexp": {
    "message":"^\\{.*?\\"product\\":\\"(toothbrush|hairspray)\\".*?\\}$"
  }
}'

And here's what I'm getting when I look at it using query(), cat(), and valid_json():

> query(qry)
Error in query(qry) : valid_json(json) is not TRUE
> cat(qry)
{
  "regexp": {
    "message":"^\{.*?\"product\":\"(toothbrush|hairspray)\".*?\}$"
  }
}
> elasticsearchr:::valid_json(qry)
[1] FALSE
attr(,"offset")
[1] 29
attr(,"err")
[1] "lexical error: inside a string, '\\' occurs before a character which it may not.\n           {   \"regexp\": { \"message\":\"^\\{.*?\\\"product\\\":\\\"(toothbrush|\n                     (right here) ------^\n"

I need to be able to escape the internal ", {, and } because these are reserved characters according to the docs and I need to a) treat them literally, and b) not close the string they are contained within early.

[AlexIoannides]

At the risk of stating the obvious, all JSON validation is handled entirely by the validate function from the jsonlite package. I think your best bet is to take a look over there for potential solutions.

[jeremy-j-ackso]

Ah, I had not taken notice of that. Didn't dig too deeply. Thanks for the pointer!

[AlexIoannides]

Sorry I couldn't be more helpful. The jsonlite package is well documented and supported so I reckon you stand a good chance of finding a solution.