`tabular-remap` (Transform Node)

The Tabular Remap node transforms a tabular dataset (CSV or TSV) into a new format by applying column-level operations defined in its configuration.

This node is ideal for standardizing schema, renaming columns, creating derived metrics, or injecting constants before data aggregation.

Configuration Schema

Property	Type	Required	Description
`inputFormat`	`"auto" \| "csv" \| "tsv"`	✖	How to parse the input file. Defaults to `"auto"`.
`outputFormat`	`"csv" \| "tsv"`	✖	Format of the output. Defaults to `"csv"`.
`failOnMissingColumns`	`boolean`	✖	If `true`, the node fails when a `rename` source column does not exist.
`treatEmptyAsNone`	`boolean`	✖	When enabled, empty strings are interpreted as `None` in transform expressions.
`columns`	`array<object>`	✔	List of column operations (rename, transform, constant). Evaluated by ascending `priority`.

Each columns[] entry supports:

Field	Required	Description
`name`	✔	Name of the output column.
`priority`	✖	Execution order (lower values run first). Defaults to insertion order.
`action`	✔	`"rename"`, `"transform"`, or `"constant"`.
`source`	only for `rename`	Source column name to copy from.
`expr`	only for `transform`	Python expression evaluated against the input row (variable: `row`).
`value`	only for `constant`	Static value assigned to the column.

Example Configuration


{
  "inputFormat": "auto",
  "outputFormat": "csv",
  "failOnMissingColumns": true,
  "treatEmptyAsNone": true,
  "columns": [
    { "name": "timestamp", "priority": 10, "action": "rename", "source": "time" },
    { "name": "driver", "priority": 20, "action": "rename", "source": "driver_name" },
    { "name": "speed_kph", "priority": 30, "action": "transform", "expr": "float(row['speed_mps']) * 3.6 if row['speed_mps'] is not None else None" },
    { "name": "unit", "priority": 40, "action": "constant", "value": "kph" }
  ]
}

Example Input and Output

Input CSV:


time,driver_name,speed_mps
2025-01-01T00:00:00Z,Alice,10
2025-01-01T00:00:01Z,Bob,12.5

Output CSV:


timestamp,driver,speed_kph,unit
2025-01-01T00:00:00Z,Alice,36.0,kph
2025-01-01T00:00:01Z,Bob,45.0,kph

Transform Expression Context

Transform expressions execute within a safe, sandboxed Python environment:

Access the current row as row['column_name'].
Built-ins available: math, float, int, str, len, min, max, sum, abs, round.
Empty strings become None if treatEmptyAsNone is true.

Example:


{ "name": "speed_category", "action": "transform", "expr": "'fast' if float(row['speed_kph']) > 40 else 'slow'" }

Example: Mixed Operations

Configuration:


{
  "columns": [
    { "name": "C1", "priority": 10, "action": "rename", "source": "a" },
    { "name": "C2", "priority": 20, "action": "transform", "expr": "int(row['b']) + 5" },
    { "name": "C3", "priority": 30, "action": "constant", "value": "processed" }
  ]
}