Coming from KNIME

If you're a KNIME user, you already understand node-based workflows and the power of visual data science pipelines. Sigilweaver follows a similar paradigm with some differences in scope and approach.

Terminology Mapping

KNIME	Sigilweaver	Notes
Workflow	Workflow	Both are saved files
Node	Tool	Dragged from palette onto canvas
Connection	Wire	Links between nodes/tools
Port	Socket	Input/output connection points
File Reader	Input tool	Load CSV, Parquet, Excel
File Writer	Output tool	Save data to files
Column Filter	Select tool	Choose columns to keep/remove
Row Filter	Filter tool	Keep rows matching condition
GroupBy	Summarize tool	Aggregate by groups
Joiner	Join tool	Combine two tables
Concatenate	Union tool	Stack tables vertically
Math Formula	Formula tool	Create calculated columns
Sorter	Sort tool	Order rows by columns

Common Tasks

Loading Data

KNIME:
Drag File Reader node → Pick file → Execute → View data

Sigilweaver:
Drag Input tool → Pick your file → Click to preview data

Filtering Rows

KNIME:
Drag Row Filter node → Choose column and condition

Sigilweaver:
Drag Filter tool → Write expression like pl.col("value") > 100 → True/False outputs

Key difference: KNIME uses GUI configuration. Sigilweaver uses Polars expressions - more concise but requires learning syntax.

Selecting Columns

KNIME:
Drag Column Filter node → Include/exclude columns

Sigilweaver:
Drag Select tool → Check columns to keep → Optionally rename, reorder, or cast

Creating Calculated Columns

KNIME:
Drag Math Formula or Column Expressions node → Build formula

Sigilweaver:
Drag Formula tool → Expression: pl.col("a") + pl.col("b") → Name output column

Aggregating Data

KNIME:
Drag GroupBy node → Select group columns → Add aggregations

Sigilweaver:
Drag Summarize tool → Set columns to "Group By" → Choose aggregation type (Sum, Mean, Count, etc.)

Joining Tables

KNIME:
Drag Joiner node → Connect top and bottom inputs → Configure join keys

Sigilweaver:
Drag Join tool → Wire left (L) and right (R) inputs → Configure keys → Three outputs: matched (J), unmatched left (L), unmatched right (R)

Sorting Data

KNIME:
Drag Sorter node → Select columns and sort order

Sigilweaver:
Drag Sort tool → Configure columns and ascending/descending order

Key Differences

1. Scope: Data Prep vs. Full Analytics

KNIME is a comprehensive analytics platform:

• Machine learning nodes (classification, clustering, etc.)
• Statistical analysis
• Database connectors
• R and Python integration
• Reporting and visualization

Sigilweaver focuses specifically on data transformation pipelines:

• Extract, transform, load (ETL)
• Data cleaning and preparation
• No ML or stats (yet)

Think of Sigilweaver as doing what KNIME does before you get to machine learning - the data wrangling and preparation phase.

2. Expression Language

KNIME uses its own expression syntax and visual builders.

Sigilweaver uses Polars expressions (Python-like):

# Filter
pl.col("age") > 18

# Formula
pl.col("price") * 1.1

# Conditional
pl.when(pl.col("status") == "active")
  .then(1)
  .otherwise(0)

See the Expressions Guide for complete documentation.

3. Execution Model

KNIME: Right-click nodes → Execute → Results stored in node → View with output connector

Sigilweaver: Lazy evaluation by default:

• Workflow is optimized before execution
• Preview data without executing entire pipeline
• Efficient with large datasets

4. Lightweight and Fast

KNIME is Java-based and can be resource-heavy. Sigilweaver is built with:

• Electron for the UI (cross-platform desktop)
• Polars (Rust) for data processing - extremely fast
• Python backend for orchestration

Sigilweaver is lighter, starts faster, and processes data efficiently.

Your First Workflow

Let's build a typical KNIME pattern: load data, filter, create a calculated column, and aggregate.

Steps:

1. Load data:

   • Drag Input tool to canvas
   • Pick your CSV file
   • Click to preview

2. Filter rows:

   • Drag Filter tool, wire from Input
   • Expression: pl.col("country") == "USA"
   • Use the T (true) output

3. Calculate discount price:

   • Drag Formula tool, wire from Filter's T output
   • Expression: pl.col("price") * 0.9
   • Output column name: discounted_price

4. Aggregate by category:

   • Drag Summarize tool, wire from Formula
   • Set category to Group By
   • Set discounted_price to Mean
   • Set product_id to Count

5. Save results:

   • Drag Output tool, wire from Summarize
   • Configure format (CSV, Parquet, Excel)
   • Execute workflow

What's Missing (Compared to KNIME)?

Sigilweaver is focused on data transformation, so it doesn't have:

• Machine learning nodes
• Statistical analysis
• Database connectors (yet - planned)
• R/Python scripting nodes
• Interactive visualizations

If you need those features, KNIME is still your best bet. But for pure data wrangling and ETL, Sigilweaver is lighter, faster, and open source.

Next Steps

• Interface Guide - Learn canvas navigation and shortcuts
• Expressions Basics - Master Polars expression syntax
• Tools Reference - Detailed docs for every tool

If you're comfortable with KNIME's node-based approach, Sigilweaver will feel familiar. The main difference is learning Polars expressions instead of KNIME's GUI builders - but expressions are more powerful and faster to use once you know them.