How to use the aggregate command with the MongoDB connector in Matlab?

I have just started using MongoDB with Matlab (R2021b). Currently, I am using the mongoc class (https://ch.mathworks.com/help/database/ug/mongoc.html) to connect to my MongoDB database.

However, after going through the documentation I am unable to find the aggregate method equivalent to the command in MongoDB. Is is possible to run aggregation operations with the other available methods (find, insert, update, remove) or is there a raw db.runCommand() method as described in the documentation (https://docs.mongodb.com/manual/reference/command/)? Am I constrained with the find method for querying?

Any help would be greatly appreciated!

Answer 1

[Shubham]

Hi William,

As of MATLAB R2021b, the direct support for MongoDB's aggregation framework through the mongoc class is not explicitly documented. The mongoc class primarily provides support for basic CRUD (Create, Read, Update, Delete) operations via its find, insert, update, and remove methods. This might seem limiting if you're looking to perform more complex queries and aggregations directly through MATLAB's built-in MongoDB interface.

However, there are a couple of approaches you can consider to execute aggregation queries or use MongoDB's more advanced features from MATLAB:

• For some aggregation needs, you might be able to use the find method with MongoDB's $where operator for simple aggregations or filtering. This is more limited than the full aggregation pipeline but can be useful for straightforward queries. This method you also have mentioned in your query.
• Write your aggregation queries in a MongoDB shell script or a JavaScript file and execute them using the MongoDB command-line interface (CLI) or a system call from MATLAB. You can then read the results back into MATLAB for further processing. For eg:

command = 'mongo database --eval "db.collection.aggregate([...])" > outputFile.json';
[status, cmdout] = system(command);
% Then read outputFile.json into MATLAB

• Depending on your specific needs and environment, you might consider using a MongoDB driver for a programming language supported by MATLAB (e.g., Java, Python) via MATLAB's interoperability features. For example, you can use Python's pymongo library within MATLAB to perform aggregation queries and then process the results in MATLAB.
• If the direct MongoDB support in MATLAB does not meet your needs, another approach is to connect to MongoDB using a JDBC or ODBC driver. This method requires setting up a JDBC or ODBC connection to MongoDB, which might involve additional steps and tools (like MongoDB's BI Connector for ODBC connections). Once set up, you can use MATLAB's database toolbox to execute arbitrary SQL queries, which might include SQL representations of aggregation operations.

While MATLAB's mongoc class provides a straightforward way to interact with MongoDB for basic operations, performing aggregation queries directly is not explicitly supported as of R2021b. Using external scripts, leveraging other programming languages' MongoDB drivers through MATLAB, or setting up a JDBC/ODBC connection are potential workarounds to execute complex MongoDB operations, including aggregations, from MATLAB.