Avoid Multidimensional Arrays and Nested Lists

ID

java.avoid_multidim_collection

Severity

low

Remediation Complexity

trivial

Remediation Risk

low

Remediation Effort

low

Resource

Efficiency

Language

Java

Tags

best-practice, efficiency

Description

Reports multidimensional arrays (e.g., int[][]) and nested list types (e.g., List<List<X>>). Two parallel single-dimension arrays are more memory-efficient and cache-friendly than a multidimensional array or nested collection structure.

Rationale

A Java multidimensional array like int[][] is an array of array references, not a contiguous block of memory. Each inner array is a separate heap object with its own header and possible different GC timing. This indirection harms CPU cache locality and increases GC overhead. Nested lists compound the problem with additional object-per-element overhead.

// Bad: multidimensional array -- array of array references
int[][] matrix = new int[100][100];

// Bad: nested list -- heavy object overhead
List<List<Integer>> grid = new ArrayList<List<Integer>>();

Remediation

Use parallel single-dimension arrays or a flat array with index arithmetic. For small fixed-size structures, consider dedicated data classes.

// Good: flat array with index arithmetic
int[] matrix = new int[100 * 100];
int value = matrix[row * 100 + col];

// Good: parallel arrays
int[] xCoords = new int[100];
int[] yCoords = new int[100];

References