[GrandComicsDatabase/gcd-django] Critical Data Contamination: Reprint credits displacing original creators in gcd_story_credit (Issue #710)
ivanwine
ivanwine created an issue (GrandComicsDatabase/gcd-django#710)Description:
While #531 and #546 deal with the UI of migration, the actual result in current SQL dumps is that "hollow" records are being created where original creators are lost and replaced by reprint/localization staff. This makes the normalized credit table unreliable for API consumers and third-party tools in its current state.
Bug Report: Original Creators Missing/Overwritten in gcd_story_credit (Story 2714655)
I am developing an application using the GCD SQL dump and have identified a critical failure mode in the current state of credit normalization where foreign localization data is effectively "poisoning" the primary story record.
The Issue:
In certain newer issues (e.g., Batman/Superman #1, 2019), the legacy inline fields on gcd_story (script, pencils, etc.) are blank. However, the normalized gcd_story_credit table does not contain the original creators (Joshua Williamson/David Marquez). Instead, it has been populated with localization staff from foreign reprints (Dutch and Bulgarian) assigned to primary roles like 'script' and 'pencils'.
Evidence (from SQL Dump):
Querying story_id 2714655 in the gcd_story_credit table returns the following (schema: modified 2024-02-15):
credit_type_id: 1 (script) -> Linda van Bruggen (Dutch Translator)
credit_type_id: 2 (pencils) -> Веселин Чакъров (Bulgarian Artist)
credit_type_id: 3 (inks) -> Веселин Чакъров
credit_type_id: 4 (colors) -> Aart Cornelissen
credit_type_id: 5 (letters) -> John J. Hill (Original Letterer)
The Problem: The original writer (Williamson) and artist (Marquez) are entirely absent from the normalized table for this story_id. Because the legacy gcd_story fields are also blank for this record, the original creative team is effectively "erased" from the database for any tool relying on these fields.
The Impact:
There appears to be no flag in gcd_story_credit to distinguish "Original Publication" credits from "Reprint/Localization" credits. When the inline legacy fields are wiped during migration, and foreign credits are the only ones present in the normalized table, there is currently no programmatic way to retrieve or identify the correct original creative team.
Questions: 1. Is there a planned schema update to include an is_original flag, a language_id, or a direct issue_id link in the gcd_story_credit table to prevent this "contamination"?
2. Until such a flag exists, what is the recommended way to filter out localization staff when querying credits for an original English edition?
—
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications on the go with GitHub Mobile for iOS or Android.
You are receiving this because you are subscribed to this thread.
JochenGCD
—
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications on the go with GitHub Mobile for iOS or Android.
You are receiving this because you are subscribed to this thread.
JochenGCD
jochengcd left a comment (GrandComicsDatabase/gcd-django#710)StoryCredit refers to CreatorNameDetail, not Creator
https://www.comics.org/issue/2011849/ has the right credits., so does the dump.
You are mixing creator_name_detail_id with creator_id
https://www.comics.org/creator_name/14239/sequences/
https://www.comics.org/creator/14239/
—
Reply to this email directly, view it on GitHub, or unsubscribe.
Triage notifications on the go with GitHub Mobile for iOS or Android.
You are receiving this because you are subscribed to this thread.