Skip to main content

Transformers

Transformers are pipeline nodes that clean, reshape, or enrich tabular data. A runnable workflow starts with a Data Input, passes data through one or more transformers, and can finish with a Data Output.

Transformer Categories​

The public catalog organizes transformers by purpose:

  • Combine & Shape: Merge or union data from multiple inputs, or reshape it with pivot and unpivot operations.

  • Clean & Format: Change text case, filter text with regular expressions, drop columns, remove duplicate rows, and clean text values.

  • Transform & Enhance: Rename, encode, normalize, or split columns, and fill missing values with Impute.

  • Detect & Manage: Detect or handle outliers, infer schema types, and change data types.

  • Organize: Sort rows by one or more columns.

Reusable Pipeline Steps​

Transformers make data-cleaning and preparation rules reusable. A chain of transformers shows the processing order clearly and applies the same rules each time the pipeline runs.

Getting Started with Transformers​

  1. Add and configure a Data Input node. Choose its dataset and sampling settings.
  2. Choose the transformations you need from the Transformer References.
  3. Add each transformer, connect the nodes, and configure its parameters.
  4. Add a Data Output node if you want to download a file or write the result to a configured cloud or Snowflake destination.

Use Chat Commands to ask the AI to build a pipeline, analyze a result, plan the work, or delegate a complex request.

note

The Transformer References describe each node's inputs, settings, and output.