Snowpark Migration Accelerator: Conversion¶

The Snowpark Migration Accelerator (SMA) always starts by creating an assessment, but it doesn’t stop there. The SMA can also convert some elements of a Spark workload to run in Snowpark.

The following topics will be covered in this section.

• How the conversion works

• Conversion Quick Start

• Understanding the Conversion Summary

• Conversion Output (reports, logs, but most importantly, the code)

• Working with the Output Code

• Next Steps

It is important to keep in mind that the SMA will only convert some references from the Spark API to the Snowpark API. It will not convert them all. The SMA team understands this and has built the SMA with that in mind. The SMA puts in issue and error codes into the SMA to better inform the user what they need to do next. This is perhaps the most valuable part of a conversion in the SMA: it converts what it can, and reports on what it can’t.

As a result, when doing any conversion, it’s essential to understand how to interpret the issues output by the SMA. There is an entire section devoted just to issue analysis including:

• Approach to resolving issues

• Issue codes by Source

• Troubleshooting through the issues

• Dealing with Workarouands

• Deploying the output code

All of this will be essential to understand if a conversion is going to be successful.

Time to jump in with conversion. Let’s start with how it works.