Check Your Spark Code

Whether your code runs on LakeSail is a question about your code, not a general claim. Three steps answer it: scan the code, port one real job, and look up anything the scan flags in the Sail docs.

1. Scan your code

Point the Spark-to-Sail scanner at your code before you change anything. Paste a snippet, upload a file, or run it over a whole repository. It reports which of the Spark features your code uses are supported by Sail.

A scan tells you which calls are supported. It does not tell you that the output matches on your data. That is step 2.

2. Port one job and diff

1. Pick a job that uses the patterns you care about: your common joins, your UDFs, your write path.
2. Run it on LakeSail as a job with the file or SQL as it is, or from your own client through a session.
3. Run the same job where it runs today.
4. Compare the output: row counts, schema, and a content hash.

Do this on at least one real workload before moving anything else. A clean diff on a representative job covers the functions, data, and edge cases your code hits.

3. Look up what the scan flags

If the scan flags something, the Sail docs say what is supported and what to change:

• Migrating from Spark
• SQL support
• DataFrame API support
• Functions
• UDFs
• Delta Lake
• Iceberg

Most flagged items have a short path: a Scala or Java UDF becomes a Python UDF, a Hive-specific function gets swapped. Code built on the RDD API does not run on Sail and needs a rewrite on DataFrames or SQL.

Next

• Quickstart: set up the platform and run your first job.
• Introduction: what LakeSail is and what you can run on it.