Skip to content

DATA_ACCESS.md

Latest commit

 

History

History
31 lines (22 loc) · 1.29 KB

DATA_ACCESS.md

File metadata and controls

31 lines (22 loc) · 1.29 KB
Raw

Data access

KDD Cup 2012 Track 2 data are not redistributed in this repository. Obtain access independently through the official Kaggle competition page and comply with its rules.

Place the downloaded archive at:

<repo>/data/external/track2.zip

The archive used for this project had exactly 2,882,914,995 bytes and local SHA-256 E1FBB172130BA9AE55593F74E5B6BFDD74E1E6046A0B540FFB6B07DD087F5686. This is the checksum of this project's verified official Kaggle download; it is not represented as a Kaggle-published checksum.

Expected members include:

  • track2/training.txt
  • track2/queryid_tokensid.txt
  • track2/purchasedkeywordid_tokensid.txt
  • track2/titleid_tokensid.txt
  • track2/descriptionid_tokensid.txt
  • track2/userid_profile.txt

Verify, ingest by streaming the ZIP member, and build the DuckDB marts:

.\.venv\Scripts\python.exe scripts�erify_kdd_source.py
.\.venv\Scripts\python.exe scripts\ingest_kdd.py
.\.venv\Scripts\python.exe scripts\ingest_kdd_companions.py
.\.venv\Scripts\python.exe scripts�uild_kdd_marts.py

The parser enforces the official 12-field TSV schema. It never requires loading the 10.6 GB uncompressed training member into memory. Generated Parquet and DuckDB files are ignored by Git.