Relational

Relational methods are useful for permissions checks, validation mapping, and UI logic where you need to know how two paths relate — whether one is a prefix of another, whether they overlap, or what structural relationship they have.

Mental model

Think of every path as a location in the data tree. A shorter path like profile covers a wider subtree; a deeper path like profile.name points at a specific spot inside that subtree. startsWith and covers are direction-reversed views of the same containment:

deeperPath.startsWith(widerPath) — does the deeper location fall under the wider one?
widerPath.covers(deeperPath) — does the wider location contain the deeper one?

(covers is not like Array.prototype.includes; it asks about data-tree containment, not element membership.)

startsWith

Returns true if the path begins with every segment of other.

import { path } from "data-path";

type User = { profile: { name: string; age: number } };

const namePath = path((u: User) => u.profile.name);
const profilePath = path((u: User) => u.profile);

namePath.startsWith(profilePath)   // true  — "profile.name" starts with "profile"
profilePath.startsWith(namePath)   // false

The argument can be a Path, a lambda, or a { segments } object.

covers

Returns true if this path’s location covers other — i.e. this path is a prefix of other (the inverse direction of startsWith).

profilePath.covers(namePath)   // true  — "profile" covers "profile.name"
namePath.covers(profilePath)   // false

Wildcards and direction

When one path contains wildcards (from .each() or .deep()), use whichever side reads most naturally:

Called on	Argument	Returns `true` when
template	`.covers(concrete)`	template’s pattern covers the concrete path
concrete	`.startsWith(template)`	concrete falls under the template prefix

interface R { items: Array<{ name: string }> }

const allNames = path((r: R) => r.items).each((i: { name: string }) => i.name);
const one = path((r: R) => r.items[0].name);

allNames.covers(one)      // true — the template covers this concrete path
one.startsWith(allNames)  // true — the concrete falls under the template prefix

The two are logically equivalent — pick the form that reads naturally for your use case.

For ** (deep wildcard), any number of intermediate segments can match:

interface Root { tree: { a: { b: { c: string } } } }

const deep = path((r: Root) => r.tree).deep();      // ["tree", "**"]
const leaf = path((r: Root) => r.tree.a.b.c);       // ["tree", "a", "b", "c"]

leaf.startsWith(deep)   // true — "**" matches "a.b.c"
deep.covers(leaf)       // true

equals

Returns true if both paths have identical segments.

const p1 = path((u: User) => u.profile.name);
const p2 = path((u: User) => u.profile.name);

p1.equals(p2)   // true
p1.equals(profilePath)   // false

match

Returns a MatchResult describing the structural relationship, or null if there is none.

type MatchResult = {
  relation: "equals" | "parent" | "child" | "covers" | "covered-by";
};

namePath.match(profilePath)  // { relation: "child" }   — name is under profile
profilePath.match(namePath)  // { relation: "parent" }  — profile contains name
namePath.match(namePath)     // { relation: "equals" }

const unrelated = path((u: User) => u.profile.age);
namePath.match(unrelated)    // null — no relationship

Relation meanings:

Relation	Meaning
`"equals"`	Both paths are identical
`"parent"`	`this` is a prefix of `other`
`"child"`	`other` is a prefix of `this`
`"covers"`	`this` contains a wildcard that matches `other`
`"covered-by"`	`other` contains a wildcard that matches `this`

Practical example — permission guard

type Store = { users: User[]; settings: { theme: string } };

const settingsPath = path((s: Store) => s.settings);

function canWrite(targetPath: Path<Store, unknown>): boolean {
  // Allow writes only outside the settings subtree
  return !targetPath.startsWith(settingsPath);
}

canWrite(path((s: Store) => s.users[0].name))    // true
canWrite(path((s: Store) => s.settings.theme))   // false