6DammK9/25110801-229M

---
license: mit
---

# 25110801-229M #

- [Apache Parquet](https://en.wikipedia.org/wiki/Apache_Parquet) *Dataset* from a MySQL *Database* dump.

- [Author's article.](https://webbhk.substack.com/about)

- [The Repository.](https://drive.google.com/drive/folders/13mzAvvufXC3QmH8OvMVx2juUi7qdnVLL)

- [One of the ongoing site.](https://webbsite.0xmd.com/dbpub/) and [video](https://youtu.be/xV2RV3z4URc?si=u48eWYjcPbEgmxpV).

- Data cutoff: **251101** (it keeps updating as on 251108)

- *I will update more contents when I think it is more appropriate. Sorry for no further elaboration.*

- *Carpe diem!*

## Contents inside ##

- `parquet`: Raw data exported from MySQL. Data tables only. Empty tables are absent.

- `parquet_uuid`: `parquet` with UUID for each row.

- `parquet_fk`: Foreign Keys linking 2 UUIDs under `parquet_uuid`.

- `dump`: "The riddle". *This folder triggered HF and blocked by upload, so I used rename.*

## How to use ##

- Bring a Machine with ~~64GB~~ 128GB of RAM. 64GB for `parquet_*`, 112GB for `dump`. However the original MySQL only takes 6GB for *single user*.

- *Good luck guessing this riddle. However I think it is quite obvious if you know the theories well. It is just a bit abstract, general, and simple.*

## Future Plan ##

- Source code... I'll rename this repo when I really want to promote it. ~~The (dataset) codes will be in Github. Server code is likely "never exist".~~ See `static`. I made it pure frontend.

- [Diagram of the dataset conversion](https://github.com/6DammK9/nai-anime-pure-negative-prompt/blob/main/ch97/parquet_neo4j.md) are located in Github whatsever.

## Hints on "the riddle" ##

- [News article but not in English (0293).](https://www.am730.com.hk/column/%E8%B2%A1%E7%B6%93/%E4%BA%A4%E5%8F%89%E6%8C%81%E8%82%A1%E6%98%AF%E9%82%AA%E6%83%A1%E9%82%84%E6%98%AF%E5%96%84%E8%89%AF%E4%B9%8B%E6%A7%8D-/234671)

- [Another news article (0273).](https://www.hkcd.com.hk/content/2017-12/22/content_1074882.html)

```cypher
//Need to flip all the "sholders_holder" first
MATCH (a:sholdings)-[rel:sholders_holder]->(b:persons)
WITH rel
CALL apoc.refactor.invert(rel)
YIELD input, output
RETURN output

//Enjoy
MATCH (v00:stocklistings{StockCode:'0273',StockExID:'1'})
WITH v00
CALL apoc.path.spanningTree(v00, {
relationshipFilter: 'stocklistings_issue>|FK_issue_orgs>|FK_organisations_persons>|sholders_holder>|sholders_issue>'
})
YIELD path
WITH nodes(path) as list
CALL apoc.nodes.cycles(list, {
relTypes: [
"stocklistings_issue",
"FK_issue_orgs",
"FK_organisations_persons",
"sholders_holder",
"sholders_issue"
]
})
YIELD path AS p00
WITH p00, nodes(p00) AS v01
MATCH p01=(:stocklistings)-[stocklistings_issue]->()-[FK_issue_orgs]->(v02:organisations)
WHERE v02 IN v01
RETURN p00, p01 LIMIT 1000
```

## Visualising / quick start for "the riddle" ##

- Host the entire `static` as a mini web server.

- Read codes and change the auth content (e.g. password).

- Use `?stockcode=0273&limit=1000` to change the query parameters.

- *You will find that the code is suprisingly low. It is (really) abstract, general, and simple.*