Skip to main content

Formula

Create or replace calculated columns using visual operation builders or Polars expressions.

Sockets​

SocketDirectionDescription
inputInputData to transform
outputOutputData with new/updated column(s)

Configuration​

Each Formula tool can contain multiple operations - add as many as you need in a single node. Operations run sequentially, so later operations can reference columns created by earlier ones.

Each operation has two modes:

ModeDescription
Basic (default)Visual builder with dropdowns - no coding required
CustomWrite Polars expressions directly for advanced logic
No coding required

The Basic mode provides 50+ built-in operations across math, text, dates, type casting, null handling, and conditional logic. Most data transformations can be done entirely through dropdowns and parameter inputs - no expression syntax needed.

Basic Mode (Visual Builder)​

The visual builder walks you through each operation step by step:

  1. Pick a source column from the dropdown
  2. Choose an operation - the list filters to show only operations compatible with your column's data type
  3. Set parameters - each operation shows relevant inputs (values, column pickers, format strings, etc.)
  4. Name the output column - a sensible default is generated automatically, or enter your own

An expression preview at the bottom shows the generated Polars code.

Available Operations​

CategoryOperations
MathAdd, Subtract, Multiply, Divide, Modulo, Power, Round, Floor, Ceiling, Absolute Value, Square Root, Logarithm, Clip, Negate, Percent of Total, Cumulative Sum, Rank
TextUppercase, Lowercase, Title Case, Trim, Replace, Substring, Pad, String Length, Contains, Starts With, Ends With, Extract (Regex), Split & Get, Concatenate Columns
Date/TimeExtract Year/Month/Day/Hour/Minute/Second, Day of Week, Truncate, Format Date, Date Difference
Type CastingCast Type, Parse Date from String, Parse DateTime from String
Null HandlingFill Null (Value), Fill Null (Strategy: forward/backward/mean/min/max/zero), Coalesce, Is Null, Is Not Null
ConditionalIf/Then/Else, Map/Replace Values
OtherRow Number, Literal Value, Shift (Lag/Lead)

Multiple Operations​

A single Formula node can hold multiple operations:

  • Click Add Operation to add more
  • Drag and drop to reorder
  • Collapse/expand individual operations to keep things tidy
  • Duplicate an operation to create a variation
  • Each operation can independently use Basic or Custom mode

Custom Mode (Advanced)​

Switch any operation to Custom mode to write Polars expressions directly. This is useful when you need logic that goes beyond the built-in operations.

FieldRequiredDescription
Target ColumnYesSelect existing column to replace, or choose "(New Column)"
New Column NameIf newName for the new column (only when creating new)
Formula ExpressionYesPolars expression defining the column value

Creating a New Column​

In Basic mode, select a source column and operation - the output column name is generated automatically. In Custom mode:

  1. Select "(New Column)" from the Target Column dropdown
  2. Enter a name for the new column
  3. Write the expression

Replacing an Existing Column​

In Basic mode, the output column name defaults to the source column - just leave it as-is to replace in place. In Custom mode:

  1. Select the column name from the Target Column dropdown
  2. Write the expression

Expression Examples​

Arithmetic​

# Calculate total from quantity and unit price
pl.col("quantity") * pl.col("unit_price")

# Calculate profit margin
(pl.col("revenue") - pl.col("cost")) / pl.col("revenue") * 100

# Add 10% markup
pl.col("price") * 1.1

String Operations​

# Combine first and last name
pl.col("first_name") + pl.lit(" ") + pl.col("last_name")

# Convert to uppercase
pl.col("status").str.to_uppercase()

# Extract first 3 characters
pl.col("code").str.head(3)

Conditional Logic​

# Categorize by value
pl.when(pl.col("amount") >= 1000)
  .then(pl.lit("High"))
  .when(pl.col("amount") >= 100)
  .then(pl.lit("Medium"))
  .otherwise(pl.lit("Low"))

# Flag condition
pl.when(pl.col("status") == "active")
  .then(pl.lit(True))
  .otherwise(pl.lit(False))

Date Operations​

# Extract year
pl.col("date").dt.year()

# Calculate age in days
(pl.lit(pl.date(2024, 1, 1)) - pl.col("birth_date")).dt.total_days()

Type Casting​

# Convert string to integer
pl.col("id_string").cast(pl.Int64)

# Convert to string
pl.col("numeric_code").cast(pl.Utf8)

Null Handling​

# Replace nulls with a default
pl.col("value").fill_null(0)

# Replace nulls with another column
pl.col("primary").fill_null(pl.col("fallback"))

Notes​

  • Column naming: If you create a column with the same name as an existing column, it replaces that column
  • Expression must be valid: Invalid expressions show an error in configuration
  • Output type: The formula determines the output column's data type automatically
  • Operation chaining: Operations within a Formula node run sequentially - operation 2 can reference a column created by operation 1
  • Mode mixing: Each operation in a multi-operation Formula can independently use Basic or Custom mode