data.map

data.map transforms fields on every row. Each entry in fields defines an output column name and a template expression that computes its value. Expressions reference input columns, apply arithmetic, call string functions, or use conditionals.

Input: Table Output: Table

Basic config

nodes:
  compute-total:
    type: data.map
    config:
      include_original: true
      fields:
        total: "{{ price * quantity }}"
        label: "{{ category | upper }}"

Config reference

Field	Type	Required	Description
fields	map	yes	Output column name to template expression
include_original	boolean	no	Keep all input columns alongside mapped fields. Default: false

When include_original is true, the output contains every input column plus the mapped fields. If a mapped field name matches an existing column, the mapped value replaces it in place. When false, the output contains only the mapped fields.

Expression syntax

Expressions use {{ ... }} delimiters. Inside them, you reference columns by name and apply operations.

Column references

fields:
  email_copy: "{{ email }}"

Arithmetic

fields:
  total: "{{ price * quantity }}"
  margin: "{{ revenue - cost }}"
  tax: "{{ subtotal * 0.19 }}"

String concatenation

Multiple {{ }} blocks in one expression produce string concatenation:

fields:
  full_name: "{{ first_name }} {{ last_name }}"
  greeting: "Hello {{ first_name }}!"

Pipe filters

Filters transform values. Syntax: {{ field | filter }}.

Filter	Result type	Description
lower	string	Lowercase
upper	string	Uppercase
trim	string	Strip whitespace
length	number	String length
round	number	Round to nearest integer
ceil	number	Round up
floor	number	Round down
abs	number	Absolute value
to_string	string	Cast to string
to_number	number	Cast to number (null on failure)
strip_currency	number	Remove currency symbols, parse as number
extract_number	number	Pull first numeric value from string
parse_date_dmy	date	Parse dd/mm/yyyy
parse_date_mdy	date	Parse mm/dd/yyyy
parse_date_ymd	date	Parse yyyy-mm-dd

fields:
  email_lower: "{{ email | lower }}"
  price_cents: "{{ price_text | strip_currency }}"
  joined: "{{ join_date | parse_date_dmy }}"

Filters with arguments

replace and regexp_replace take arguments:

fields:
  clean_phone: '{{ phone | replace("+", "00") }}'
  no_tags: '{{ body | regexp_replace("<[^>]+>", "") }}'

Coalesce (null fallback)

fields:
  display_name: '{{ nickname ?? "Anonymous" }}'

Conditional (ternary)

fields:
  tier: '{{ score > 0.7 ? "high" : "low" }}'

Quoted identifiers

For column names with special characters, use double quotes inside the expression:

fields:
  amount: '{{ "Debit/Credit" | to_number }}'

Examples

Computed columns with originals

nodes:
  enrich-orders:
    type: data.map
    config:
      include_original: true
      fields:
        total: "{{ price * quantity }}"
        discount_amount: "{{ total * discount_pct }}"
        status_label: '{{ status == "paid" ? "Paid" : "Pending" }}'

Projection (select + transform)

nodes:
  clean-contacts:
    type: data.map
    config:
      fields:
        name: "{{ first_name }} {{ last_name }}"
        email: "{{ email | lower | trim }}"
        score: "{{ raw_score | round }}"

Since include_original defaults to false, the output contains only name, email, and score.

Type casting

nodes:
  parse-amounts:
    type: data.map
    config:
      include_original: true
      fields:
        amount: "{{ amount_text | strip_currency }}"
        quantity: "{{ qty_string | to_number }}"
        label: "{{ id | to_string }}"

Casts use TRY_CAST internally — unparseable values become null instead of errors.