R : Updating an entry in mongodb using mongolite
Asked Answered
G

1

8

I have a mongo database with information that I am passing to some R scripts for analysis. I am currently using the mongolite package to pass the information from mongo to R.

I have a field in each mongo entry called checkedByR, which is a binary that indicates whether the entry has been analysed by the R scripts already. Specifically, I am collecting a mongo entry by its respective mongo ID, running the scripts on the entry, assigning the checkedByR field with a 1, and then moving on.

For completeness, I am querying the database with the following request:

library(mongolite)

mongoID <- "1234abcd1234abcd1234"

m <- mongolite::mongo(url = "mongodb://localhost:27017",
                      collection = "collection",
                      db = "database")

rawData <- m$find(query = paste0('{"_id": { "$oid" : "',mongoID,'" }}'), 
                  fields = '{"_id" : 1, 
                             "checkedByR" : 1, 
                             "somethingToCheck" : 1}')

checkedByR <- 1

However, I am having trouble successfully updating the mongo entry with the new checkedByR field.

I realise that an update function exists in the mongolite package (please consider : https://cran.r-project.org/web/packages/mongolite/mongolite.pdf), but I am having trouble gathering relevant examples to help me complete the updating process.

Any help would be greatly appreciated.

Ghyll answered 4/6, 2016 at 15:30 Comment(0)
Y
5

the mongo$update() function takes a query and a update argument. You use the query to find the data you want to update, and the update to tell it which field to update.

Consider this example

library(mongolite)
    
## create some dummy data and insert into mongodb
df <- data.frame(id = 1:10,
  value = letters[1:10]
)
    
mongo <- mongo(collection = "another_test", 
  db = "test", 
  url = "mongodb://localhost")
    
mongo$insert(df)
    
## the 'id' of the document I want to update
mongoID <- "575556825dabbf2aea1d7cc1"
    
## find some data
rawData <- mongo$find(query = paste0('{"_id": { "$oid" : "',mongoID,'" }}'), 
  fields = '{"_id" : 1, 
  "id" : 1, 
  "value" : 1}'
)
    
## ...
## do whatever you want to do in R...
## ...

## use update to query on your ID, then 'set' to set the 'checkedByR' value to 1

mongo$update(
  query = paste0('{"_id": { "$oid" : "', mongoID, '" } }'),
  update = '{ "$set" : { "checkedByR" : 1} }'
)

## in my original data I didn't have a 'checkedByR' value, but it's added anyway

Update

the rmongodb library is no longer on CRAN, so the below code won't work


And for more complex structures & updates you can do things like

library(mongolite)
library(jsonlite)
library(rmongodb)  ## used to insert a non-data.frame into mongodb
    
## create some dummy data and insert into mongodb
lst <- list(id = 1,
  value_doc = data.frame(id = 1:5,
  value = letters[1:5],
  stringsAsFactors = FALSE),
  value_array = c(letters[6:10])
)
    
## using rmongodb
mongo <- mongo.create(db = "test")
coll <- "test.another_test"
    
mongo.insert(mongo, 
  ns = coll, 
  b = mongo.bson.from.list(lst)
)
    
mongo.destroy(mongo)

## update document with specific ID
mongoID <- "5755f646ceeb7846c87afd90"
    
## using mongolite
mongo <- mongo(db = "test", 
  coll = "another_test", 
  url = "mongodb://localhost"
)
    
    
## to add a single value to an array
mongo$update(
  query = paste0('{"_id": { "$oid" : "', mongoID, '" } }'),
  update = '{ "$addToSet" : { "value_array" :  "checkedByR"  } }'
)
    
## To add a document  to the value_array
mongo$update(
  query = paste0('{"_id": { "$oid" : "', mongoID, '" } }'),
  update = '{ "$addToSet" : { "value_array" : { "checkedByR" : 1} } }'
)
    
## To add to a nested array
mongo$update(
  query = paste0('{"_id": { "$oid" : "', mongoID, '" } }'),
  update = '{ "$addToSet" : { "value_doc.value" :  "checkedByR" } }'
)
    
rm(mongo); gc()

see mongodb update documemtation for further details

Yettayetti answered 6/6, 2016 at 11:2 Comment(0)

© 2022 - 2024 — McMap. All rights reserved.