Export the two optimised Eurostat extracts as CSV, straight from the live RINF knowledge graph — sections of line with their track parameters, or the TEN-T / corridor classification flags. Both queries are the tested versions in eurostat/: they use the predicates the graph actually carries today, aggregate multi-valued fields instead of multiplying rows, and can be run per country or across the whole EU.
Every date is ISO 8601, YYYY-MM-DD. In the CSV it is plain text, but a spreadsheet opening the CSV
converts it to its own local format — except dates before 1900, such as the 1800-01-01 placeholder France uses
for “since always”, which stay as typed — so one column ends up in two formats. The .xlsx keeps every date as
YYYY-MM-DD text, so they all read the same; lengths, coordinates and flags stay numbers.