decode (Compute)

When To Use

Use decode to reverse base64 or url encoding on a single column, row by row, leaving the other columns untouched. Reach for it when an API or command returns encoded values — a Base64-encoded name, description, or token payload, or percent-encoded identifiers extracted from URLs — and you need the original text for matching, translation tables, or display attributes in mapping.

Syntax

sources:
  sensors:
    type: http
    method: get
    path: /api/v1/sensors
    resultContent: body
    computes:
    - type: json2Csv
      entryKey: /sensors
      properties: /id;/encodedName;/status
    # Column 3 (/encodedName) holds a Base64-encoded display name
    - type: decode
      column: 3
      encoding: base64

Properties

Table Transformation Example

With column: 3 and encoding: base64 (columns shown after json2Csv, whose leading entry column is column 1):

With encoding: url, a value such as sensor%20bay%201 becomes sensor bay 1.

Recommended Pattern

• Decode as early as possible, right after the payload-to-table compute, so every later compute (keepOnlyMatchingLines, translate, replace) and mapping works on readable text.
• Pick the encoding from the API documentation or an observed sample; a wrong choice yields garbage, not an error you can grep for.
• Combine with extract or substring when only part of a column is encoded: isolate the encoded fragment in its own column first, then decode it.

Common Mistakes

• Decoding the wrong column after an upstream compute changed the table shape (indexes are 1-based and shift when columns are added or removed).
• Forgetting the leading entry column produced by json2Csv: the first JSON property sits in column 2, so the decoded field is one column further right than expected.
• Applying base64 decoding to plain text: the result is binary garbage that silently corrupts the row.
• Confusing direction: decode recovers original text; use encode^[1] to produce Base64/URL-encoded values.

Community Examples