Datasets:

---

OPheno

/

Weed-phenology

like 0

Follow

OPheno 3

I would leave crop & targetCode columns to be as close to the source as possible. Future us will thank us.

@JorgeAEP let's check column of csv and make sure we are matching there.

Another reason was: The columns "crop (EPPO)" and "weed (target) eppo" have spaces and parentheses, which causes friction in Python (df["crop (EPPO)"] instead of df.crop_eppo), in DuckDB/SQL queries, and in Parquet round-trips.

Common naming conventions:

• snake_case — country_code, seeding_date. Standard in Python (PEP 8), Ruby, Rust, SQL, and most data/CSV contexts. Also called lower_snake_case when you need to disambiguate.
• SCREAMING_SNAKE_CASE — COUNTRY_CODE. Used for constants in most languages, and for some legacy database column conventions.
• camelCase — countryCode, trialNumber. Standard in JavaScript, Java, and Swift. First word lowercase, subsequent words capitalized. Note that your trialNumber column is actually camelCase, inconsistent with the rest of your schema.
• PascalCase (or UpperCamelCase) — CountryCode. Used for class names in most languages, and for types in C#/TypeScript.
• kebab-case (or dash-case, lisp-case) — country-code. Used in URLs, CSS, HTML attributes, and Lisp dialects. Invalid in most programming language identifiers because the dash parses as a minus operator.
• Train-Case — Country-Code. Rare; sometimes seen in HTTP headers (Content-Type is technically this).
• flatcase — countrycode. No separation. Rare and hard to read.

Its preferred to stick with source so you don't need to change it every time a new dataset is published.