I'm failing miserably to be able to restore a single collection into an existing database.
I'm running Ubuntu 14.04 with mongo version 2.6.7
There is a dump/mydbname/contents.bson based off my home directory.
If I run
mongorestore --collection contents --db mydbname
Then I get:
connected to: 127.0.0.1
don't know what to do with file [dump]
If I add in the path
mongorestore --collection contents --db mydbname --dbpath dump/mydbname
Then I get
If you are running a mongod on the same path you should connect to that instead of direct data file access
I've tried various other combinations, options, etc. and just can't puzzle it out, so I'm coming to the community for help!
If you want to restore a single collection then you have to specifiy the dump file of the collection. The dump file of the collection is found in the 'dump/dbname/' folder. So assuming your dump folder is in your current working directory, the command would go something like -
mongorestore --db mydbname --collection mycollection dump/mydbname/mycollection.bson
I think this is now done with the --nsInclude option:
mongorestore --nsInclude test.purchaseorders dump/
dump/ is the folder with your mongodump data, test is the db, and purchaseorders is the collection.
https://docs.mongodb.com/manual/reference/program/mongorestore/
Steps to restore specific collection in the mongodb.
1) Go to the directory where your dump folder exists.
2) Execute following command by modifying according to your db name and your collection name.
mongorestore --db mydbname --collection mycollection dump/mydbname/mycollection.bson
If you get Failed: yourdbname.collection.name: error creating indexes for collection.name: createIndex error: The field 'safe' is not valid for an index specification error, then you can use following command:
mongorestore --db mydbname --collection mycollection dump/mydbname/mycollection.bson --noIndexRestore
If you are restoring multiple collections, you can use a loop:
for file in "$HOME/mongodump/dev/<your-db>/"* ; do
if [[ "$file" != "*metadata*" && "$file" != "system.*" && "$file" != "locks.*" ]]; then
file="$(basename "$file”)"
mongorestore \
--db cdt_dev \
--collection "${file%.*}" \ # filename w/o extension
--host "<your-host>" \
--authenticationDatabase "<your-auth-db>" \
-u "user" \
-p "pwd" \
"$HOME/mongodump/dev/<your-db>/$file"
fi;
done
How do I export the results of a MongoDB command to a flat file
For example, If I am to get db.collectionname.find() into a flat file.
I tried db.collectionname.find() >> "test.txt" doesnt seem to work.
you can try the following from the command line
mongo 127.0.0.1/db --eval "var c = db.collection.find(); while(c.hasNext()) {printjson(c.next())}" >> test.txt
assuming you have a database called 'db' running on localhost and a collection called 'collection' this will export all records into a file called test.txt
If you have a longer script that you want to execute you can also create a script.js file
and just use
mongo 127.0.0.1/db script.js >> test.txt
I hope this helps
I know of no way to do that from the mongo shell directly, but you can get mongoexport to execute queries and send the results to a file with the -q and -o options:
mongoexport -h mongo.dev.priv -d models -c profiles -q '{ $query : { _id : "MRD461000" } }' -o MRD_Series1.json
The above hits queries the profiles collection in the models database grabbing the JSON document for _id = "MRD641000". Works for me.
Use this
mongo db_name --username user --password password < query1.js >> result.txt
Try this - returns a json file with the data of the query, you can change .json for .txt and other.
mongoexport --db products --collection clicks --query '{"createdInt":{$gte:20190101}, "clientId":"123", "country":"ES"}' --out clicks-2019.json
Having missed the db needing to be the actual db in Peshkira's answer, here is a general syntax for a one liner in shell (assuming no password):
mongo <host>:<db name> --eval "var x = <db name>.<collection name>.<query>; while(x.hasNext()) { printjson( x.next() ) }" >> out.txt
I tested it both on my mac and Google cloud Ubuntu 15 with Mongo 3+.
Install MongoDB Compass, then it will have a tool to export query result to Json/CSV files.
mongoexport --host 127.0.0.1 --port 27017 --username youruser -p yourpass \
-d yourDatabaseName -c collectionName --type csv \
--fields field1,field2 -q '{"field1" : 1495730914381}' \
--out report.csv
mongoexport --db db_name --collection collection_name --csv --out file_name.csv -f field1,field2, field3
CSV file with contact information:
Name,Address,City,State,ZIP
Jane Doe,123 Main St,Whereverville,CA,90210
John Doe,555 Broadway Ave,New York,NY,10010
Running this doesn't add documents to the database:
$ mongoimport -d mydb -c things --type csv --file locations.csv --headerline
Trace says imported 1 objects, but in the MongoDB shell running db.things.find() doesn't show any new documents.
What am I missing?
Your example worked for me with MongoDB 1.6.3 and 1.7.3. Example below was for 1.7.3. Are you using an older version of MongoDB?
$ cat > locations.csv
Name,Address,City,State,ZIP
Jane Doe,123 Main St,Whereverville,CA,90210
John Doe,555 Broadway Ave,New York,NY,10010
ctrl-d
$ mongoimport -d mydb -c things --type csv --file locations.csv --headerline
connected to: 127.0.0.1
imported 3 objects
$ mongo
MongoDB shell version: 1.7.3
connecting to: test
> use mydb
switched to db mydb
> db.things.find()
{ "_id" : ObjectId("4d32a36ed63d057130c08fca"), "Name" : "Jane Doe", "Address" : "123 Main St", "City" : "Whereverville", "State" : "CA", "ZIP" : 90210 }
{ "_id" : ObjectId("4d32a36ed63d057130c08fcb"), "Name" : "John Doe", "Address" : "555 Broadway Ave", "City" : "New York", "State" : "NY", "ZIP" : 10010 }
I was perplexed with a similar problem where mongoimport did not give me an error but would report importing 0 records. I had saved my file that didn't work using the OSX Excel for Mac 2011 version using the default "Save as.." "xls as csv" without specifying "Windows Comma Separated(.csv)" format specifically. After researching this site and trying the "Save As again using "Windows Comma Separated (.csv)" format, mongoimport worked fine. I think mongoimport expects a newline character on each line and the default Mac Excel 2011 csv export didn't provide that character at the end of each line.
We need to execute the following command:
mongoimport --host=127.0.0.1 -d database_name -c collection_name --type csv --file csv_location --headerline
-d is database name
-c is collection name
--headerline If using --type csv or --type tsv, uses the first line as field names. Otherwise, mongoimport will import the first line as a distinct document.
For more information: mongoimport
you will most likely need to authenticate if you're working in production sort of environments. You can use something like this to authenticate against the correct database with appropriate credentials.
mongoimport -d db_name -c collection_name --type csv --file filename.csv --headerline --host hostname:portnumber --authenticationDatabase admin --username 'iamauser' --password 'pwd123'
I use this on mongoimport shell
mongoimport --db db_name --collection collection_name --type csv --file C:\\Your_file_path\target_file.csv --headerline
type can choose csv/tsv/json
But only csv/tsv can use --headerline
You can read more on the offical doc.
Check that you have a blank line at the end of the file, otherwise the last line will be ignored on some versions of mongoimport
When I was trying to import the CSV file, I was getting an error. What I have done.
First I changed the header line's column names in Capital letter and removed "-" and added "_" if needed. Then Typed below command for importing CSV into mongo
$ mongoimport --db=database_name --collection=collection_name --type=csv --file=file_name.csv --headerline
Robert Stewart have already answered for how to import with mongoimport.
I am suggesting easy way to import CSV elegantly with 3T MongoChef Tool (3.2+ version). Might help someone in future.
You just need to select collection
Select file to import
You can also unselect data which is going to import. Also many options are there.
Collection imported
See how to import video
First you should come out of the mongo shell and then execute the mongoimport command like this:
Manojs-MacBook-Air:bin Aditya$ mongoimport -d marketdata -c minibars
--type csv
--headerline
--file '/Users/Aditya/Downloads/mstf.csv'
2017-05-13T20:00:41.989+0800 connected to: localhost
2017-05-13T20:00:44.123+0800 imported 97609 documents
Manojs-MacBook-Air:bin Aditya$
Robert Stewart's answers is great.
I'd like to add that you also can type your fields with --columHaveTypes and --fields like this :
mongoimport -d myDb -c myCollection --type csv --file myCsv.csv
--columnsHaveTypes --fields "label.string(),code.string(),aBoolean.boolean()"
(Careful to not have any space after the comma between your fields)
For other types, see doc here : https://docs.mongodb.com/manual/reference/program/mongoimport/#cmdoption-mongoimport-columnshavetypes
For the 3.4 version, please use the following syntax:
mongoimport -u "username" -p "password" -d "test" -c "collections" --type csv --file myCsv.csv --headerline
After 3 days, I finally made it on my own. Thanks to all the users who supported me.
My requirement was to import the .csv (with no headline) to remote MongoDB instance. For mongoimport v3.0.7below command worked for me:
mongoimport -h <host>:<port> -u <db-user> -p <db-password> -d <database-name> -c <collection-name> --file <csv file location> --fields <name of the columns(comma seperated) in csv> --type csv
For example:
mongoimport -h 1234.mlab.com:61486 -u arpitaggarwal -p password -d my-database -c employees --file employees.csv --fields name,email --type csv
Below is the screenshot of how it looks like after import:
where name and email are the columns in the .csv file.
Given .csv file I have which has only one column with no Header, below command worked for me:
mongoimport -h <mongodb-host>:<mongodb-port> -u <username> -p <password> -d <mongodb-database-name> -c <collection-name> --file file.csv --fields <field-name> --type csv
where field-name refers to the Header name of the column in .csv file.
C:\wamp\mongodb\bin>mongoexport --db proj_mmm --collection offerings --csv --fieldFile offerings_fields.txt --out offerings.csv
Just use this after executing mongoimport
It will return number of objects imported
use db
db.collectionname.find().count()
will return the number of objects.
use :
mongoimport -d 'database_name' -c 'collection_name' --type csv --headerline --file filepath/file_name.csv
mongoimport -d test -c test --type csv --file SampleCSVFile_119kb.csv --headerline
check collection data:-
var collections = db.getCollectionNames();
for(var i = 0; i< collections.length; i++)
{
print('Collection: ' + collections[i]);
// print the name of each collection
db.getCollection(collections[i]).find().forEach(printjson);
//and then print the json of each of its elements
}
1]We can save xsl as .csv file
2] Got to MongoDB bin pathon cmd - > cd D:\Arkay\soft\MongoDB\bin
3] Run below command
> mongoimport.exe -d dbname -c collectionname --type csv --file "D:\Arkay\test.csv" --headerline
4] Verify on Mongo side using below coomand.
>db.collectioname.find().pretty().limit(1)
Strangely no one mentioned --uri flag:
mongoimport --uri connectionString -c questions --type csv --file questions.csv --headerline
Sharing for future readers:
In our case, we needed to add the host parameter to make it work
mongoimport -h mongodb://someMongoDBhostUrl:somePORTrunningMongoDB/someDB -d someDB -c someCollection -u someUserName -p somePassword --file someCSVFile.csv --type csv --headerline --host=127.0.0.1
Make sure to copy the .csv file to /usr/local/bin or whatever folder your mondodb is in
All these answers above are great. And the way to go on a full featured application.
But if you want to prototype fast, want flexibility as the collection still changes as well as to minimize your early code base, there is a much simpler way that is not much discussed.
You can basically forego mongoimport by now. I could have saved 3 hours if it was mentioned here on this question. So let me share for others:
Mongodb has a GUI called Mongo Compass has both csv and json import features out of the box in a matter of clicks. It is an official part of the Mongo ecosytem. At the time of writing it is free and it works very well for my use case.
https://www.mongodb.com/products/compass
You simply get MongoDB compass running on your machine by following the simple installation. A couple of fields for DB connection and authentication directly in the GUI.
Import the csv/json file. It took less than a second on a 30KB file to be parsed before user (me) validates.
Validate the "type" of each property. Great feature, I could directly mention the property types such as booleans, integers, etc. In my experience, they seem all default to string. You can update before importing. Dates were more finicky and needed special attention on the coding side.
One click further the csv is a collection in your mongo db local or on the cloud. Voila!
If you have multiple files and you want to import all of them using python, you can do the following.
import os
import subprocess
# directory of files
dir_files = 'C:\data'
# create list of all files
_, _, fns = next(os.walk(dir_files))
files = [os.path.join(dir_files, fn) for fn in fns]
# mongotool address
mongotool = r'C:\Program Files\MongoDB\Server\4.4\bin\mongoimport.exe'
# name of mongodb database
mydatabase = 'mydatabase'
# name of mongodb collection
mycollection = 'mycollection'
# import all files to mongodb
for fl in files:
commands =[mongotool, '--db', mydatabase,
'--collection', mycollection,
'--file', fl,
'--type', 'tsv',
'--headerline']
subprocess.Popen(commands, shell=True)