MongoDB to Delta Lake

This page provides you with instructions on how to extract data from MongoDB and load it into Delta Lake. (If this manual process sounds onerous, check out Stitch, which can do all the heavy lifting for you in just a few clicks.)

What is MongoDB?

MongoDB, or just Mongo, is an open source NoSQL database that stores data in JSON format. It uses a document-oriented data model, and data fields can vary by document. MongoDB isn't tied to any specified data structure, meaning that there's no particular format or schema for data in a Mongo database.

What is Delta Lake?

Delta Lake is an open source storage layer that sits on top of existing data lake file storage, such AWS S3, Azure Data Lake Storage, or HDFS. It uses versioned Apache Parquet files to store data, and a transaction log to keep track of commits, to provide capabilities like ACID transactions, data versioning, and audit history.

Getting data out of MongoDB

The process of pulling data out of MongoDB depends on how you've loaded data into MongoDB. In some cases, it may be impossible to extract all of your data, because NoSQL databases don't require structure (i.e. specific columns). Relational databases, such as those used for data warehouses, use a more traditional, rigid structure. You'll need to defined a structure in the relational database into which you can insert MongoDB data.

Don't stress about the confusing data structure. Lots of the data that's loaded into MongoDB is created by a computer, so it probably has a pretty predictable structure. If you can find specific fields that exist for every record, you're well on your way. Make sure these fields appear in the records of each collection you'd like to replicate from MongoDB. There are many ways to do this. The most popular method to get data from MongoDB is to use the find() command.

Sample MongoDB data

MongoDB stores and returns JSON-formatted data. Here's an example of what a response might look like to a query against the products collection.

db.products.find( { qty: { $gt: 25 } }, { _id: 0, qty: 0 } )

{ "item" : "pencil", "type" : "no.2" }
{ "item" : "bottle", "type" : "blue" }
{ "item" : "paper" }

Loading data into Delta Lake on Databricks

To create a Delta table, you can use existing Apache Spark SQL code and change the format from parquet, csv, or json to delta. Once you have a Delta table, you can write data into it using Apache Spark's Structured Streaming API. The Delta Lake transaction log guarantees exactly-once processing, even when there are other streams or batch queries running concurrently against the table. By default, streams run in append mode, which adds new records to the table. Databricks provides quickstart documentation that explains the whole process.

Keeping MongoDB data up to date

Fine job! You are the proud developer of a script that moves data from MongoDB to your data warehouse. This works as a one-shot deal. It's good to think about what will happen when there is new and updated data in MongoDB.

One option that works would be to load the entire MongoDB dataset all over again. That would certainly update the data, but it's not very efficient and can also cause terribly latency.

The smartest way to get data updated from MongoDB would be to identify keys that can be used as bookmarks to store where you script left off on the last run. Fields like updated_at, modified_at, or other auto-incrementing data are useful here. With that done, you can set up your script as a cron job or continuous loop to identify new data as it appears.

Other data warehouse options

Delta Lake on Databricks is great, but sometimes you need to optimize for different things when you're choosing a data warehouse. Some folks choose to go with Amazon Redshift, Google BigQuery, PostgreSQL, or Snowflake, which are RDBMSes that use similar SQL syntax, or Panoply, which works with Redshift instances. Others choose a data lake, like Amazon S3. If you're interested in seeing the relevant steps for loading data into one of these platforms, check out To Redshift, To BigQuery, To Postgres, To Snowflake, To Panoply, and To S3.

Easier and faster alternatives

If all this sounds a bit overwhelming, don’t be alarmed. If you have all the skills necessary to go through this process, chances are building and maintaining a script like this isn’t a very high-leverage use of your time.

Thankfully, products like Stitch were built to move data from MongoDB to Delta Lake automatically. With just a few clicks, Stitch starts extracting your MongoDB data, structuring it in a way that's optimized for analysis, and inserting that data into your Delta Lake data warehouse.