SALT: Sales Autocompletion Linked Business Tables Dataset

SALT is a dataset of linked business tables for sales autocompletion, presented at the NeurIPS'24 Table Representation Workshop. It was created by SAP and integrated into the RelBench framework in 2025. The dataset is designed for machine learning tasks involving structured business data.

Use Cases

• Train models for sales autocompletion based on linked business tables.
• Benchmark link prediction algorithms on structured, multi-table business data.
• Research table representation learning for enterprise applications.

Strengths

• Dataset is integrated into the RelBench framework, suggesting a standardized evaluation setup.
• Associated with a peer-reviewed paper presented at NeurIPS'24.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset scale are unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

SAP

Freshness

Last updated 2025-08-07 22:34:16; freshness should be verified.