Validation

This page outlines the step-by-step process for the validation procedure.

To proceed with the validation process, follow the steps outlined below:

1. Copy the snowpark-checkpoints-output folder, generated during the collection process, into the validation workload.

Copying collection process output into the validation workload

1. Open the validation workload in VS Code to begin the validation process.

Validation Workload

1. Generate checkpoints using the checkpoints.json file.

To generate checkpoints you can do one of the following actions:

• Generate them by accepting the suggesting message:

Load found checkpoints message

• Execution "Snowflake: Load All Checkpoints command"

Load All Checkpoints Command

Once all checkpoints are loaded, your files should appear as follows:

File with checkpoints

1. Run the Python file to execute the checkpoints validation process.

When running a python file that contains validation checkpoints, the validation results are going to be shown in the copied "snowpark-checkpoints-output" folder as "checkpoints_validation_results.json":

The "checkpoints_validation_results.json" contains the unified results of the collection process

{
    "results": [
        {
            "checkpoint_name": "sample$BBVOC7$df1$1",
            "file": "sample.py",
            "line_of_code": 10,
            "result": "PASS",
            "timestamp": "2025-05-05T15:32:29.248917"
        },
        {
            "checkpoint_name": "sample$BBVOC7$df2$1",
            "file": "sample.py",
            "line_of_code": 12,
            "result": "PASS",
            "timestamp": "2025-05-05T15:32:31.137536"
        },
        {
            "checkpoint_name": "sample$BBVOC7$df3$1",
            "file": "sample.py",
            "line_of_code": 17,
            "result": "PASS",
            "timestamp": "2025-05-05T15:32:33.133002"
        }
    ]
}

The validation results, as seen above, will contain the comparison result between the PySpark and Snowpark DataFrames.