Bridging Generation and Training: A Systematic Review of Quality Issues in LLMs for Code

TL;DR

This systematic review explores how training data quality issues influence code generation failures in large language models, proposing a taxonomy, causal framework, and discussing detection and mitigation techniques.

Contribution

It introduces a unified taxonomy and causal framework linking training data issues to code quality problems in LLMs, and reviews current detection and mitigation strategies.

Findings

Training data imperfections significantly impact code quality in LLMs.

Shift from post-generation filtering to proactive data governance.

Identified open challenges and future research directions.

Abstract

Large language models (LLMs) frequently generate defective outputs in code generation tasks, ranging from logical bugs to security vulnerabilities. While these generation failures are often treated as model-level limitations, empirical evidence increasingly traces their root causes to imperfections within the training corpora. Yet, the specific mechanisms linking training data quality issues to generated code quality issues remain largely unmapped. This paper presents a systematic literature review of 114 primary studies to investigate how training data quality issues propagate into code generation. We establish a unified taxonomy that categorizes generated code quality issues across nine dimensions and training data quality issues into code and non-code attributes. Based on this taxonomy, we formalize a causal framework detailing 18 typical propagation mapping mechanisms. Furthermore, we synthesize state-of-the-art detection and mitigation techniques across the data, model, and generation lifecycles. The reviewed literature reveals a clear methodological shift: quality assurance is transitioning from reactive, heuristic-based post-generation filtering toward proactive, data-centric governance and closed-loop repair. Finally, we identify open challenges and outline research directions for developing reliable LLMs for code through integrated data curation and continuous evaluation. Our repository is available at https://github.com/SYSUSELab/From-Data-to-Code.