Choose the right distribution

Download files from the exact Zenodo version record. The record includes file sizes, checksums, methods, schemas and a changelog. Consult those materials before choosing a large distribution.

An exact release citation#

Ukolov, D. (2026). MODAVIS Pipe Organ Dataset (POD)
(Version 1.6.0) [Data set]. Zenodo.
https://doi.org/10.5281/zenodo.22308263

The dataset contribution is distributed under CC BY 4.0. Original sources, linked media and software retain their applicable rights; the dataset license does not grant unrestricted rights to those materials.

Choose the smallest adequate starting point#

Begin with the research question and the required relationships. A map-oriented projection can be convenient for geographic browsing; an account-level analysis may require the public core or a research distribution. Check the documented profile rather than assuming that a smaller database contains every table and qualification of the full release.

Inspect a downloaded SQLite database without changing it#

from pathlib import Path
import sqlite3

path = Path("your-downloaded-database.sqlite").resolve()
# Read-only mode fails if the file does not exist.
with sqlite3.connect(path.as_uri() + "?mode=ro", uri=True) as db:
    rows = db.execute(
        "SELECT name FROM sqlite_schema "
        "WHERE type = ? ORDER BY name", ("table",)
    )
    for (name,) in rows:
        print(name)

This first inspection deliberately makes no assumptions about table names. Compare the result with the release’s schema and profile documentation before writing domain queries. Keep the downloaded input unchanged and write analysis outputs elsewhere. If you unpack a compressed file, distinguish the download checksum from the identity of the unpacked database.

For semantic and server-based work#

Use the complete N-Triples distributions when the analysis requires the full chosen mapping profile. Catalogue summaries are lighter discovery products and do not establish complete graph coverage. A server projection can support a different execution environment while still requiring a pinned release, documented restore and explicit query method.

• Record the version DOI and exact filenames before processing.
• Check the repository file information and the release’s checksum inventory.
• Read schema, methods, profile descriptions and known limitations together.
• Save queries, exclusions and output identities alongside the analysis.