Rmongodb检测和清除不良文件

Question

I've got a mongo collection that sometimes got messed up by using mongo.insert() instead of mongo.insert.batch() on bson objects that were created with mongo.bson.from.df().

Please see this reproducible example, first I introduce a bad document followed by several good documents and I've tried to detect and remove the bad documents. The detection seems to be successful, but not the removal with mongo.remove() by _id.

If this script worked, length(x) should be one record shorter after the detection and removal.

Thanks.

library(rmongodb)
ns <- "testdb.del"
mongo <- mongo.create()
mongo.drop(mongo,ns)
df1 <- data.frame(numbers=rnorm(7),alphas=letters[1:7],monumbs=1:7)
df1
b <- mongo.bson.from.df(df1)
mongo.insert(mongo,ns,b) #insert bad document
mongo.insert.batch(mongo,ns,b) #insert good documents
x <- mongo.find.all(mongo,ns)
length(x)

for(i in 1:length(x)){
  if(is.na(names(x[[i]]["numbers"]))){ #bad documents don't have the names in the right place
    print("bad document. Trying to remove.")
    rm.id <- mongo.bson.from.list(x[[i]]["_id"]) #grab _id of bad document and turn to bson
    mongo.remove(mongo,ns,rm.id)  #remove it.
  }
}

x <- mongo.find.all(mongo,ns)
length(x)

Answer 1

Use x <- mongo.find.all(mongo,ns, mongo.oid2character = F) . In your query mongodb oid s were converted into characters.