Transformations are data manipulation operations that you configure in the mapping and normalization steps of Supplier Data Manager (SDM). They let you clean and reformat supplier-provided values — for example, standardizing text capitalization, removing extra whitespace, applying regex substitutions, padding short codes to a fixed length, or scaling numeric values by a factor.

Transformations are defined in the step's JSON configuration under a "transformations" key and applied to field values automatically during data processing.

Types of transformations

1. Capitalize

Adjusts the capitalization of text values.

Parameters:

• name: capitalize

• which: Determines what to capitalize. Options:

  • "start": Capitalizes only the first character of the string.
  • "sentences": Capitalizes the first character of each sentence (delimited by ., ?, or !). Default.
  • "words": Capitalizes the first character of each word.
  • "all": Converts the entire string to uppercase.
  • "none": Converts the entire string to lowercase.

• mode: Determines how capitalization is applied. Options:

  • "strict": Lowercases all characters before capitalizing (equivalent to Python's str.capitalize()).
  • "loose": Only uppercases the target character, leaving the rest unchanged. Default.

Example:

> "hello world. how are you?" # (which="sentences", mode="loose")
# Output: "Hello world. How are you?"

2. Trim Whitespaces

Removes excess whitespace from text values.

Parameters:

• name: trim_whitespaces
• multiple: If true, also collapses multiple consecutive space characters within the string to a single space. If false (default), only removes leading and trailing whitespace.

Example:

> "  Hello   world!  " # (multiple=true)
# Output: "Hello world!"

3. Regex

Applies a regular expression substitution to the value.

Parameters:

• name: regex
• pattern: The regex pattern to match.
• substitution: The string to replace matched patterns with.
• flags: List of regex flags to apply. Options: "ascii", "multiline", "ignorecase", "dotall". These correspond to Python regex flags.
• force: If true, the field value is set to empty when the pattern does not match. If false (default), the value is left unchanged when there is no match.

Example:

> "Hello123World" # (pattern="\\d+", substitution="", flags=["ignorecase"])
# Output: "HelloWorld"

// Simple digits removal:
{
  "transformations": [
    {
      "name": "regex",
      "pattern": "\\d+",
      "substitution": "",
      "flags": ["ignorecase"],
      "force": false
    }
  ]
}
// Input: "Hello123World" → Output: "HelloWorld"

// Extracting part of a number sequence:
{
  "transformations": [
    {
      "name": "regex",
      "pattern": "^(\\d{8})(\\d{2})$",
      "substitution": "\\1",
      "flags": ["ignorecase"],
      "force": false
    }
  ]
}
// Input: "1234567890" → Output: "12345678"

// Removing line breaks:
{
  "transformations": [
    {
      "name": "regex",
      "flags": ["multiline"],
      "force": false,
      "pattern": "\\r?\\n",
      "substitution": ""
    }
  ]
}
// Input: "Hello\nWorld" → Output: "HelloWorld"

4. Pad

Pads a string value to a specified total length by adding a fill character on the left or right.

Parameters:

• name: pad
• fill: The character to use for padding.
• direction: Where to add the padding. Options: "left" (default) or "right".
• length: The desired total length of the string after padding. If the string is already equal to or longer than length, it is returned unchanged.

Example:

> "123" # (fill="0", direction="left", length=5)
# Output: "00123"

Common use case: EAN codes exported from Excel often lose leading zeros. For example, 0003542147089 becomes 3542147089. Setting fill to "0", direction to "left", and length to 13 restores the missing zeros.

5. Multiply

Multiplies a numeric value by a specified factor. This is useful for unit conversions (for example, converting meters to millimeters) or applying fixed coefficients.

Parameters:

• name: multiply
• factor: The number by which the input value is multiplied. Can be an integer or a decimal. Non-numeric values are left unchanged.

Example:

> 10.5 # (factor=100)
# Output: 1050.0

Usage notes

• Transformations run on all fields sent to the step on the initial pass. When a user modifies a field in the UI and submits the form, transformations run only on the fields that were modified.
• Transformations are applied in the order they are listed in the configuration.
• Transformations only apply to the data type they support: text transformations (capitalize, trim_whitespaces, regex, pad) skip non-string values, and multiply skips non-numeric values.