Zod Array Validation

Zod Array Validation

Validate arrays, tuples, records, maps, and sets with Zod's collection primitives

When to Use

• Validating lists of items with consistent shape (e.g., API response arrays, batch inputs)
• Validating key-value maps where keys and values have known types
• Working with fixed-length, typed positional data using tuples
• Ensuring collections meet cardinality constraints (non-empty, length limits)

Instructions

1. Use z.array() for lists of items with a uniform element schema:

import { z } from 'zod';

const TagsSchema = z.array(z.string());
const NumberListSchema = z.array(z.number().positive());
const UserListSchema = z.array(UserSchema);

2. Constrain array length with .min(), .max(), and .length():

const AtLeastOneSchema = z.array(z.string()).min(1, 'At least one item required');
const BoundedSchema = z.array(z.number()).min(1).max(10);
const ExactSizeSchema = z.array(z.string()).length(3, 'Must have exactly 3 items');

3. Use .nonempty() as a type-aware alternative to .min(1) — the inferred type becomes a non-empty tuple:

const NonEmptyTags = z.array(z.string()).nonempty('At least one tag is required');
type NonEmptyTags = z.infer<typeof NonEmptyTags>; // [string, ...string[]]

4. Use z.tuple() for fixed-length arrays with positional types:

const PointSchema = z.tuple([z.number(), z.number()]);
// [x, y] — exactly two numbers

const RGBSchema = z.tuple([
  z.number().min(0).max(255),
  z.number().min(0).max(255),
  z.number().min(0).max(255),
]);

// Tuple with rest element (fixed prefix + variable suffix)
const AtLeastTwoStrings = z.tuple([z.string(), z.string()]).rest(z.string());

5. Use z.record() for key-value maps where keys are strings and values share a type:

// Record with string keys and number values
const ScoreMapSchema = z.record(z.number());
// Equivalent to { [key: string]: number }

// Record with validated string keys
const EnvVarsSchema = z.record(z.string().min(1));

// Record with typed keys (uses z.enum or z.string())
const FeatFlagsSchema = z.record(z.enum(['featureA', 'featureB', 'featureC']), z.boolean());

6. Use z.map() for JavaScript Map objects:

const StringToNumberMap = z.map(z.string(), z.number());
// Parses: new Map([['a', 1], ['b', 2]])

7. Use z.set() for JavaScript Set objects:

const UniqueTagsSchema = z.set(z.string());
const BoundedSetSchema = z.set(z.number()).min(1).max(5);
// Parses: new Set([1, 2, 3])

8. Use .element to access the element schema of an array for composition:

const ItemsSchema = z.array(z.object({ id: z.string(), name: z.string() }));
const ItemSchema = ItemsSchema.element; // z.ZodObject<...>
type Item = z.infer<typeof ItemSchema>;

Details

Array vs tuple — when to choose:

Unique element validation:

Zod does not have built-in uniqueness checking. Use .refine():

const UniqueStringsSchema = z.array(z.string()).refine((arr) => new Set(arr).size === arr.length, {
  message: 'Array must contain unique values',
});

Transforming arrays:

// Sort after parsing
const SortedNumbersSchema = z.array(z.number()).transform((arr) => [...arr].sort((a, b) => a - b));

// Deduplicate
const DeduplicatedSchema = z.array(z.string()).transform((arr) => [...new Set(arr)]);

// Flatten
const FlatTagsSchema = z.array(z.array(z.string())).transform((arr) => arr.flat());

Paginated response pattern:

function paginatedSchema<T extends z.ZodTypeAny>(itemSchema: T) {
  return z.object({
    data: z.array(itemSchema),
    total: z.number().int().nonnegative(),
    page: z.number().int().positive(),
    pageSize: z.number().int().positive(),
  });
}

const PaginatedUsersSchema = paginatedSchema(UserSchema);
type PaginatedUsers = z.infer<typeof PaginatedUsersSchema>;

Source

https://zod.dev/api#arrays

Process

1. Read the instructions and examples in this document.
2. Apply the patterns to your implementation, adapting to your specific context.
3. Verify your implementation against the details and edge cases listed above.

Harness Integration

• Type: knowledge — this skill is a reference document, not a procedural workflow.
• No tools or state — consumed as context by other skills and agents.

Success Criteria

• The patterns described in this document are applied correctly in the implementation.
• Edge cases and anti-patterns listed in this document are avoided.