Data Pipelines

FScript is a strong fit for data reshaping work because the language defaults line up with transformation pipelines:

values are immutable
array helpers are curried
pipes keep the flow left to right

Example: active user names

import Array from 'std:array'
import String from 'std:string'

type User = {
  id: String,
  name: String,
  active: Boolean,
}

displayNames = (users: User[]): String[] => users
  |> Array.filter((user) => user.active)
  |> Array.map((user) => String.uppercase(user.name))

Why this style works well

Each step does one thing:

Array.filter narrows the collection
Array.map reshapes the values
the pipe makes the order read like a recipe

Because the APIs are data-last, partial application is natural:

onlyActive = Array.filter((user: User) => user.active)
toUpperName = Array.map((user: User) => String.uppercase(user.name))

displayNames = (users: User[]): String[] => users
  |> onlyActive
  |> toUpperName

Comparison to JavaScript

A JavaScript version would likely use users.filter(...).map(...). FScript chooses imported helpers instead so the behavior is explicit and consistent across the language.

Tips for good pipelines

keep each pipeline stage small
extract named helpers when a lambda gets busy
perform validation near the boundary before entering the pure pipeline
return new records instead of trying to “patch” old ones in place