MongoDB GridFS

Introduction to MongoDB GridFS

Working with MongoDB, we often have to tackle the need of storing large files, these files can be images, video and audio files, etc. This is where MongoDB’s GridFS, also known as Grid File System, can be used. Basically, dividing such large files into small chunks and storing each chunk in a separate document, is the main purpose of GridFS. This Grid File System is largely used to store files that cannot be stored in a collection while maintaining the limit. The MongoDB’s GridFS also has limitations like the files stored in GridFS can have an impact of your working files saved in collections. Performance can be affected due to storage of large files.

Syntax:

Syntax is an important aspect of any executable command. Here we have to see one notable thing, that the query to insert a file into gridfs is not executed through mongodb shell, but instead from a Linux, Windows or Mac command prompt.

Given below is the standard syntax for inserting a simple file in GridFS.

mongofiles -d database-name put filename

The above syntax will simply insert a file into a database. The “put” does the job of insertion, while database name and file name are destination and source. Along with put, there are multiple options for deleting, viewing in the form of list and removing the files.

How GridFS works in MongoDB?

MongoDB has a limit for each document size that can be stored, now the large files cannot be stored single handedly due to file size, this is where GridFS can be used. MongoDB’s GridFS works by saving a multiple number of chunks, of a single large file. This way, the large files can be stored and retrieved easily when needed. The usual recommended size for each separate chunk is 256kb.

Every file that exceeds the limit of 16 MB, which is for BSON document size limit, that file will be stored and retrieved using GridFS. It is also important to know that GridFS does not support Atomic Updating. So, to put it simply. GridFS is a file system, that stores data, of larger size, but in mongodb collections and in form of chunks.

Examples of MongoDB GridFS

Given below are the examples mentioned:

Example #1

We will experiment our very first example with a simple audio file. Right now we don’t have any specific collection for gridfs files, but as we know, MongoDB has an amazing feature of creating new collections when we insert a document into non existing collection.

Following is the query for inserting a file into gridfs, unlike other insert operations where we insert files through Mongo Shell, here we will insert to Machine Terminal.

Code:

/usr/bin$ mongofiles -d gridfs put /home/sulaksh/Desktop/potkesd/c.mp3

The above query is a standard format for inserting files into gridfs. We have navigated the terminal to the location where mongofiles, which is executable file is saved. Our query begins with the keyword, mongofiles, followed by the Storage Options code, “-d”. This -d specifies the database where the file is to be stored. This -d option is followed by the database name, here our database is named as gridfs. We, at the moment don’t have any database with such name, but it’ll be created.

Output:

Now, how to retrieve the data that we saved using mongofiles is simply by executing find query.

Code:

db.fs.files.find().pretty()

The above query will simply return all documents stored in the fs.files collection, and adding pretty() makes it represent the data in JSON format.

Output:

As you can see in the above screenshot, we have our filename, chunk size and other details of the insertion operation.

Example #2

Upon insertion, we understood how the operation to insert a file in GridFS works, let us now delete the same file and see how Deletion works.

Code:

/usr/bin$ mongofiles -d gridfs delete /home/sulaksh/Desktop/potkesd/c.mp3

The major difference between the put and delete query is the delete keyword itself.

Output:

Example #3

Insert and Delete are important operation just like Search. So for our next example we will implement search option along with mongofiles, in order to search for a file in gridfs.

Code:

/usr/bin$ mongofiles -d gridfs search Poster

Above query will search for any and all documents that store file which contains the Poster word.

Output:

And as you can see, it returned the record, that has a file with Poster as file name.

Example #4

We will execute a list options, which will return every file stored in the database, here it won’t return the file in chunks or parts but as whole.

Code:

mongofiles -d gridfs list

Output:

Just like other queries, the -d denotes the database and gridfs is the name of database and list of the command. Upon execution, all the present files stored in the gridfs database in the form of chunks are retrieved.

Using these put, list, search options, we can easily use and implement the Grid File System in real time application.

Conclusion

To Conclude, MongoDB’s amazing feature, GridFS, allows you to store files which exceed the size limit for a single document. Here, the files are stored in small chunks and in two different collections. One collection stores the actual data, the file while the other collection stored the meta data for the file. Put, Search, List, etc are the options provided in order to work with the Grid File System. With advantage of being able to store large files, GridFS can result in slow performance and can impact the working data set.

This is a guide to MongoDB GridFS. Here we discuss the introduction to GridFS in MongoDB, how does it work along with respective examples. You may also have a look at the following articles to learn more –

1. MongoDB Administration
2. Indexes in MongoDB
3. order by in MongoDB
4. Lookup in MongoDB