Column Name Mapping (Augur to Aveloxis)

Aveloxis uses cleaner column names internally but exposes Augur-compatible names in all materialized views for seamless 8Knot integration.

Design Philosophy

Augur’s schema uses prefixes like pr_src_*, gh_*, and pr_review_* that embed the data source into the column name. Aveloxis replaces these with descriptive names (platform_pr_id, review_state, submitted_at) since the data source is tracked separately via the data_source metadata column.

However, 8Knot and other downstream tools reference Augur’s column names directly. To maintain backward compatibility, all 20 materialized views alias their output columns to match Augur’s naming convention exactly.

The rule: Internal table columns use Aveloxis names. Materialized view output uses Augur names.

Pull Requests

Augur column	Aveloxis table column	Matview output alias
`pr_src_id`	`platform_pr_id`	`pr_src_id`
`pr_src_number`	`pr_number`	—
`pr_src_state`	`pr_state`	`pr_src_state`
`pr_src_title`	`pr_title`	—
`pr_created_at`	`created_at`	`pr_created_at`
`pr_merged_at`	`merged_at`	`pr_merged_at`
`pr_closed_at`	`closed_at`	`pr_closed_at`
`pr_augur_contributor_id`	`author_id`	`cntrb_id`
`pr_src_author_association`	`author_association`	`pr_src_author_association`
`pr_merge_commit_sha`	`merge_commit_sha`	—
`pr_body`	`pr_body`	(unchanged)

Pull Request Meta

Augur column	Aveloxis table column	Matview output alias
`pr_repo_meta_id`	`pr_meta_id`	—
`pr_head_or_base`	`head_or_base`	`pr_head_or_base`
`pr_src_meta_label`	`meta_label`	`pr_src_meta_label`
`pr_src_meta_ref`	`meta_ref`	—
`pr_sha`	`meta_sha`	—

Pull Request Reviews

Augur column	Aveloxis table column
`pr_review_author_association`	`author_association`
`pr_review_state`	`review_state`
`pr_review_body`	`review_body`
`pr_review_submitted_at`	`submitted_at`
`pr_review_src_id`	`platform_review_id`
`pr_review_node_id`	`node_id`
`pr_review_html_url`	`html_url`
`pr_review_commit_id`	`commit_id`

Pull Request Labels

Augur column	Aveloxis table column
`pr_src_description`	`label_description`
`pr_src_color`	`label_color`

Pull Request Assignees / Reviewers

Augur column	Aveloxis table column
`pr_assignee_map_id`	`pr_assignee_id`
`pr_assignee_src_id`	`platform_assignee_id`
`contrib_id`	`cntrb_id`
`pr_reviewer_map_id`	`pr_reviewer_id`
`pr_reviewer_src_id`	`platform_reviewer_id`

Issues

Augur column	Aveloxis table column
`gh_issue_id`	`platform_issue_id`
`gh_issue_number`	`issue_number`
`issue_state`	`issue_state` (unchanged)
`reporter_id`	`reporter_id` (unchanged)
`cntrb_id` (closed_by)	`closed_by_id`

Issue Events

Augur column	Aveloxis table column
`event_id`	`issue_event_id`
`issue_event_src_id`	`platform_event_id`

Issue Labels / Assignees

Augur column	Aveloxis table column
`label_src_id`	`platform_label_id`
`label_src_node_id`	`node_id`
`issue_assignee_src_id`	`platform_assignee_id`
`issue_assignee_src_node`	`platform_node_id`

Messages

Augur column	Aveloxis table column
`pltfrm_id`	`platform_id`
`msg_id`	`msg_id` (unchanged)
`cntrb_id`	`cntrb_id` (unchanged)

Repo Info

Augur column	Aveloxis table column
`stars_count`	`star_count`
`watchers_count`	`watcher_count`
`pull_request_count`	`pr_count`
`pull_requests_open`	`prs_open`
`pull_requests_closed`	`prs_closed`
`pull_requests_merged`	`prs_merged`
`committers_count`	`committer_count`

Repo Clones

Augur table/column	Aveloxis table/column
`repo_clones_data`	`repo_clones`
`repo_clone_data_id`	`repo_clone_id`
`count_clones`	`total_clones`
`clone_data_timestamp`	`clone_timestamp`

Table Names

Augur table	Aveloxis table	Notes
`augur_data.repo`	`aveloxis_data.repos`	Pluralized
`augur_data.message`	`aveloxis_data.messages`	Pluralized
`augur_data.platform`	`aveloxis_data.platforms`	Pluralized
`augur_data.*` (all others)	`aveloxis_data.*`	Same name
`augur_operations.*`	`aveloxis_ops.*`	Shortened

Schema Names

Augur schema	Aveloxis schema
`augur_data`	`aveloxis_data`
`augur_operations`	`aveloxis_ops`

Libyear Compatibility Note

Augur’s repo_deps_libyear table has a typo in the column name: current_verion (missing ‘s’). Aveloxis fixes this to current_version in the table schema, but the explorer_libyear_detail materialized view aliases the output column back to current_verion to maintain compatibility with 8Knot and any queries written against Augur’s schema.

Writing Queries

When writing queries against Aveloxis:

Against tables directly: Use Aveloxis column names (platform_pr_id, pr_state, created_at, etc.)
Against materialized views: Use Augur column names (pr_src_id, pr_src_state, pr_created_at, etc.)

The queries/ directory in the repository contains ~100 analytical SQL queries that have been rewritten from Augur’s schema to Aveloxis’s. These can be used as examples of the correct column names for direct table queries.

`pull_request_meta.meta_label` — why it is sometimes empty

meta_label is the head/base ref in GitHub’s owner:branch form (e.g. octocat:feature-x). On the GraphQL collection path (the default since v0.26.0) it is synthesized from the head/base repository’s nameWithOwner plus the ref name.

When the fork behind a PR’s head branch has been DELETED, the label is empty ('') on GraphQL-collected rows. GitHub’s REST API kept a copy of the label string on the PR object itself, so REST-era rows carry labels even for long-deleted forks; GraphQL exposes only the live repository object, which is null once the fork is gone — there is nothing to reconstruct the owner from. Measured scale on augurlabs/augur (2026-07-10): 119 of 5,263 head/base rows (~2%).

This is a deliberate honest-NULL: we do not cache or fabricate values we cannot verify (the same policy as null author_id for deleted users and null closed_by_id for deleted closers). Analysts joining on meta_label should treat '' as “fork deleted”; the branch NAME is still available in meta_ref, and the paired pull_request_repo row is likewise absent for deleted forks.

Known-empty columns (the 2026-08-19 fill-audit contract, v0.27.105)

A fleet-wide fill-rate audit of aveloxis_large (every non-empty table, type-aware populated predicates, sampled for 100M+-row tables) classified every column that was empty or near-empty in production. The classes and their canonical lists live in internal/db/column_writer_tripwire_test.go (documentedEmpty) — that test fails the build if a column is added to an audited entity table without either a writer or a documented reason, so this section and the allowlist cannot drift apart silently.

Augur schema-parity ballast (write-never, kept on purpose): the contributors.cntrb_{type,fake,lat,long,country_code,state,city,last_used} geo/classification family, issues.pull_request/pull_request_id (aveloxis stores PRs in their own table), issues.due_on, repos.repo_path, repo_labor.repo_url, pull_request_review_message_ref.pr_url/ pr_review_msg_url, pull_requests.pr_augur_contributor_id, unresolved_commit_emails.name, commits.cmt_ght_committed_at. Full Augur schema parity is a project goal; these columns stay declared and stay empty.

Absent upstream (no forge source exists): releases.updated_at (neither forge’s release payload carries it), contributors.gh_gravatar_id (GitHub deprecated the field — returns "" for everyone), issues.issue_url on GitLab (no API-URL field in the list payload), repo_info.security_audit_file (no community-profile source).

Token-scope / corpus outcomes (writer correct, data legitimately rare): releases.is_draft (public tokens never see draft releases), pull_request_repo.pr_repo_private_bool (private forks are invisible), messages.msg_sender_email (written only by the mailing-list projection — empty on every forge-sourced message), the email_message.linked_pull_request_id/linked_pr_review_id/ linked_commit_hash routing columns (writers exist; matches depend on the list corpus), contributors.gl_* (wired since v0.20.3; the fleet has one GitLab repo).

Lives on a different table (NO writer on this one — allowlisted): messages.msg_header and messages.rgls_id. Mailing-list headers and list linkage are first-class on email_message/email_message_ref; the shared message upsert never writes these two columns, so they sit in the tripwire’s documentedEmpty map — genuinely writer-less, not merely rare. (Corrected v0.27.113: an earlier revision grouped them with the writer-backed msg_sender_email, drifting from the map this section claims to mirror.)

Platform-parity notes: issue_events.action_commit_hash and pull_request_events.action_commit_hash are GitHub-only (the GitLab event mapping has no commit attribution).

Documented follow-ups (not yet implemented, deliberately):

commits.cmt_ght_committer_id — committer-identity resolution (the author twin IS resolved; the committer needs its own resolver pass and API budget).
repo_info.committer_count — derivable from facade commits data (COUNT(DISTINCT cmt_committer_email) per repo); GitHub’s mentionableUsers is not semantically equivalent.
The REST-rail efficiency refactor: both forges’ listing responses already carry labels/assignees that the per-item REST calls re-fetch (ListIssueLabels, ListPRAssignees, ListPRReviewers, …). Nearly idle on today’s fleet (GitHub runs the GraphQL rails; the GitLab fleet is one repo) — worth doing if GitLab tracking grows or REST mode runs fleet-wide during a GraphQL outage. Needs shadow-diff re-validation.

Everything else the audit flagged was FIXED in v0.27.102–v0.27.105: repos.platform_repo_id (rename-dedup identity), the org-expansion forge-id gap, repo_archived on the GraphQL→REST fallback, releases.data_source, contributor_identities.node_id/user_type on the GraphQL rail, pull_request_repo.pr_cntrb_id, pull_requests.meta_head_id/meta_base_id, repo_info.keywords (topics), repos.created_at/updated_at, and commits.cmt_whitespace (Augur-parity measurement + aveloxis rewalk-whitespace).