SQL Dataset

Overview

The SQL Dataset of ArXiv and OpenReview data provides a normalized relational database. Built on PostgreSQL, it offers efficient querying with proper foreign key relationships and maintains data integrity across entities.

Best for: Complex queries with multiple joins; transactional consistency; production applications requiring data integrity; advanced analytics with SQL.

Key Features

Normalized Schema

Platform-specific schemas (arxiv.*, openreview.*) with proper foreign key constraints and referential integrity.

Rich Relationships

Models complex academic relationships including citations, authorship, reviews, and revisions with proper cardinality.

Cross-Platform Links

Direct connections between OpenReview and ArXiv papers through openreview.arxiv mapping table.

Optimized Queries

Efficient indexing on primary keys, foreign keys, and frequently queried fields for fast data retrieval.

Database Schema

The SQL dataset represents entities and relations in separate PostgreSQL schemas. Each table corresponds to one specific entity or relation type defined in the documentation. Tables use proper primary keys and foreign key constraints.

ArXiv Schema (arxiv.*)

Table Name	Description	Primary Key	Key Columns
`arxiv.papers`	ArXiv papers with version tracking	`id`	arxiv_id (unique), base_arxiv_id, version, title, abstract, submit_date, metadata (jsonb)
`arxiv.authors`	Authors with Semantic Scholar IDs	`id`	semantic_scholar_id (unique), name, homepage
`arxiv.categories`	ArXiv subject categories	`id`	name (unique), description
`arxiv.sections`	Paper sections with full text	`id`	paper_arxiv_id (FK), content, title, appendix, section_in_paper_id
`arxiv.paragraphs`	Individual paragraphs	`id`	section_id (FK), content, paragraph_in_section_id
`arxiv.figures`	Paper figures	`id`	paper_arxiv_id (FK), path, caption, label, name
`arxiv.tables`	Paper tables	`id`	paper_arxiv_id (FK), path, caption, label, table_text

ArXiv Relationship Tables

Table Name	Description	Composite Key	Foreign Keys
`arxiv.citations`	Paper citations with context	`id`	citing_arxiv_id, cited_arxiv_id, bib_title, bib_key, citing_sections (jsonb), citing_paragraphs (jsonb)
`arxiv.paper_authors`	Paper-author relationships	paper_arxiv_id, author_id	paper_arxiv_id (FK), author_id (FK), author_sequence
`arxiv.paper_category`	Paper categorization	paper_arxiv_id, category_id	paper_arxiv_id (FK), category_id (FK)
`arxiv.paper_figures`	Paper-figure links	paper_arxiv_id, figure_id	paper_arxiv_id (FK), figure_id (FK)
`arxiv.paper_tables`	Paper-table links	paper_arxiv_id, table_id	paper_arxiv_id (FK), table_id (FK)
`arxiv.paragraph_references`	Internal figure/table references	`id`	paragraph_id (FK), paper_arxiv_id (FK), paper_section, reference_label, reference_type

OpenReview Schema (openreview.*)

Table Name	Description	Composite Key	Key Columns
`openreview.papers`	OpenReview papers	venue, paper_openreview_id	title, abstract, paper_decision, paper_pdf_link
`openreview.authors`	Authors with profiles	venue, author_openreview_id	author_full_name, email, affiliation, homepage, dblp
`openreview.paragraphs`	Paragraph-level content	venue, paper_openreview_id, paragraph_idx	section, content
`openreview.reviews`	Peer reviews	venue, review_openreview_id	replyto_openreview_id (FK), writer, title, content (jsonb), time
`openreview.revisions`	Paper revisions	venue, revision_openreview_id	original_openreview_id (FK), content (jsonb), time
`openreview.arxiv`	OpenReview-ArXiv links	venue, paper_openreview_id	arxiv_id, title

OpenReview Relationship Tables

Table Name	Description	Composite Key	Columns
`openreview.papers_authors`	Paper-author associations	venue, paper_openreview_id, author_openreview_id	All three columns (composite FK)
`openreview.papers_reviews`	Paper-review links	venue, paper_openreview_id, review_openreview_id	title, time
`openreview.papers_revisions`	Paper revision tracking	venue, paper_openreview_id, revision_openreview_id	title, time
`openreview.revisions_reviews`	Revision-review temporal links	venue, revision_openreview_id, review_openreview_id	All three columns (composite FK)

Dataset Statistics

Papers

40,000+

ArXiv & OpenReview

Paragraphs

5M+

Fine-grained content

Reviews

Comprehensive

Multiple venues

Citations

Tracked

With context

Export & Migration

Need to export data to other formats? See our format conversion guides:

Export to CSV

Export SQL tables to CSV format for pandas analysis

Format Conversion

Convert between SQL, CSV, and JSON formats