- 09:43 21 July 2020
I was experimenting with the GitHub source code and database and noticed that the database export (http://data.demozoo.org/demozoo-export.sql.gz) only has productions until Dec 2019.
There are edits, comments from 2020, but no productions.
Is that intentional or has something gone wrong with the database export?
- 11:44 22 July 2020
Thanks for the report - fixed now!
(That was a 'fun' one to track down. At the very end of 2019, I ripped out all of the old code for auto-generating screenshots for ANSI graphics, because while ANSI people *think* that would be a neat feature, when you actually give it to them they go "oh no it's all wrong, how did it not know that this one was supposed to be rendered on a green background with the Potnoodle font". The export is done by making a database dump, loading that into a separate database where we can redact sensitive things like password hashes, and dumping it out again. That initial dump file starts by dropping and recreating all of the tables. But because the next day's dump didn't have the ANSIs table in it, it also didn't have the line to drop that table. And since that table was sticking around, it couldn't drop the productions table because of the database constraint. And because that table wasn't being emptied, it couldn't load the new data into it, and ended up frozen in time on 31st December 2019. Which, given how 2020 has turned out so far, is probably not a bad place to be...)
- 07:04 23 July 2020
OMG, that does sound like a nightmare – and there was me thinking it might be something as simple as a misconfigured date range query! :)
- 21:12 23 August 2020
I'd like to do some mining on demoscene data so I was happy to find this export. However, I have no experience with web databases and SQL dumps. I just know MS Access (and the SQL statements there) and R. Is there any simple way to get the database information in csv-files?