Database design sources: can a robust logical/physical design be derived from existing legacy data sets (for example, through data profiling and reverse engineering)?

Introduction / Context:
Database design is often informed by existing data. When organizations modernize or migrate, architects perform data profiling and reverse engineering on legacy files/tables to derive entities, attributes, relationships, and constraints. This can seed a new conceptual, logical, and physical model.

Given Data / Assumptions:

• Legacy data exists in files or older databases.
• Tools can inspect data distributions, patterns, and implied keys.
• Business rules may be partially embedded in data shape/quality.

Concept / Approach:
Reverse engineering complements requirements analysis. Designers extract candidate entities, identify primary/foreign keys, detect many-to-many relationships, and choose normalization levels. Profiling highlights anomalies and informs constraints (NOT NULL, CHECK) and data types.

Step-by-Step Solution:
Profile existing data: frequency, uniqueness, outliers. Infer structure: entities, attributes, and relationships. Validate with stakeholders and refine the logical model. Map to physical structures (tables, indexes, partitions). Iterate to address data quality and performance needs.

Verification / Alternative check:
Cross-check derived design against process flows and use cases to ensure it captures business semantics, not just current data quirks.

Why Other Options Are Wrong:
Limiting to “NoSQL/OLAP/spreadsheets only” is incorrect; deriving design from data is a general practice across platforms.

Common Pitfalls:
Treating current data irregularities as permanent rules; failing to involve domain experts to confirm inferred constraints.

Final Answer:
Correct