Data types ​

WEL has 16 data types. Use the right type when data enters or leaves an expression to ensure bad values don't cause a transformation to silently fail.

_ is the input your expression receives in these examples. Refer to Input variables for more information.

Two rules govern how every type in WEL behaves:

• WEL doesn't coerce by default: Passing a number where a String is expected raises E100 rather than converting it. Nothing is silently reinterpreted to make an expression succeed, except inside string interpolation, which always converts automatically, and two Transform data flags that configure specific coercions.
• Collections never become strings by accident: A List or Map is only serialized by an explicit call such as to_json.

Both rules exist because a transformation that fails is visible, and one that silently produces the wrong value isn't.

Quick reference ​

WEL contains the following data types:

Numbers ​

WEL has three numeric types, and the choice between them is the most consequential type decision in most integrations.

Integer ​

An Integer is a whole number, with arbitrary precision. There is no 64-bit ceiling and no silent loss above 2^53.

Float ​

A Float is a binary floating-point number. Float arithmetic is fast, and appropriate for measurements, ratios, percentages, and scientific values. Float can't represent most decimal fractions exactly. Use Decimal in cases where fraction precision matters, such as when working with money.

Decimal ​

Decimal provides exact decimal arithmetic at arbitrary precision. Use it for cases where fraction precision matters, such as when working with money.

Mix numeric types ​

Integer can combine with both Decimal and Float. Decimal and Float can't directly mix:

The restriction prevents a Float from silently making a result approximate. Raising E103 forces the approximation to be explicit. Fix E103 by converting the Float operand to a Decimal, not by demoting the Decimal operand to a Float.

Text and bytes ​

The following two types hold text and raw byte data:

String ​

A String is UTF-8 text, in single or double quotes (' or ").

f"Order {_.id} has {_.count} items"

Binary ​

Binary is raw bytes, written in hexadecimal format. For example: 0x"48656C6C6F".

Dates and times ​

WEL has five temporal types. Use the type that fits your use case to avoid common date bugs.

DateTime ​

A DateTime is an exact instant with a time zone or offset. Use it to record the specific time of an event, such as the time something was created, shipped, or logged. For example: DateTime('2026-03-15T09:30:00Z').

PlainDate ​

A PlainDate is a calendar date with no time and no zone. Use it for a calendar date that's the same for everyone looking at it, such as an invoice date, a birth date, or a contract start date. For example: PlainDate('2026-03-15').

PlainTime ​

A PlainTime is a time of day with no date and no zone. Use it for a time of day with no date attached, such as a store's opening time or a daily cutoff. For example: PlainTime('09:30:00').

PlainDateTime ​

A PlainDateTime is a date and time with no zone. Use it for a wall-clock reading before its time zone is known, such as a local time a source system sends with the zone supplied separately. For example: PlainDateTime('2026-03-15T09:30:00').

Duration ​

A Duration holds a fixed clock duration, not a calendar unit, written Duration('PT2H30M'). Use it for a length of time, such as an age, an elapsed interval between two events, or a timeout.

Collections ​

WEL has two collection types. A List contains ordered data and a Map contains key-value data.

List ​

A List is an ordered collection that can contain elements of mixed type. Wrap the list in brackets and separate values with commas. For example: [1, 2, 3].

Refer to the List functions documentation for more information.

Map ​

A Map is a key-value collection and the WEL type for JSON objects. Keys are always String, and insertion order is preserved. Wrap the map in curly braces and separate pairs with commas. For example: {a: 1, b: 2}.

Refer to the Map functions documentation for more information.

Collections don't become strings ​

A List or Map never becomes a String automatically:

The restriction exists because an accidentally stringified collection, such as a whole array arriving in a destination field as [object Object] or a bracketed blob, is one of the hardest integration bugs to trace. Raising E100 forces conversions to be explicit. Call to_json for JSON, to_wel for a WEL literal, or join_to_string for a delimited list, depending on what the destination expects.

Null and Skip ​

null and Skip both describe an absence, but their behavior differs in many destination systems:

• null: The field has no value. The key is still emitted.

• Skip: Don't emit this at all. The key is left out of a map.

This distinction is especially common in CRM update APIs. Sending null clears a field, while omitting the key with Skip leaves it untouched.

Test for values ​

present? is the general test, and it treats every kind of absence the same way:

present?(false) returns true, because it's a real value a pipeline can carry deliberately, not a stand-in for missing data the way null or an empty string is.

presence reports every kind of absence as null so one fallback operator (| ) catches them all, including empty strings, whitespace-only strings, empty lists, and empty maps.

Convert between types ​

Each type has a constructor, such as Integer(...), Decimal(...), or PlainDate(...). Constructors are strict, for example: Integer('3 items') raises E102 rather than partially parsing.

The most reliable shape for a transformation is to convert every field at the start:

{
  order_id: String(_.id),
  qty: Integer(_.qty),
  total: Decimal(_.total),
  ship_date: PlainDate(_.ship_date)
}

Refer to the Conversion functions documentation for more information.

Related ​

• Standard library: Information about every WEL function.
• Conversion functions: How to construct each data type.
• Error codes: Troubleshoot errors such as E100, E101, E102, and E103.
• Transform data: Declare data types in the Transform data action.