Choosing between national and municipal data

Where a municipality publishes its own address file, you can download that or the national register extract for the same area, or both. They are different in ways that matter.

Every download page carrying more than one source shows a map you can toggle between them, with a summary of how they differ. Look at that before choosing — the differences are usually easier to see than to describe.

What each one gives you

	National register	Municipal open data
Coverage	Every province and territory	One municipality
Refresh	Twice a year	Monthly, where the publisher supports it
New construction	Appears at the next release	Often within weeks
Positional accuracy	Variable; many representative points	Usually good; the municipality assigns the address
Postal code	On every record	Frequently absent
Unit counts	Often present	Often absent
Schema stability	Consistent, documented	Varies; changes without notice
Availability	Everywhere	Where open data licensing permits

Which to choose

Take the national register when your work spans more than one municipality, involves postal codes, needs dwelling or unit counts, or has to be reproducible across releases. Consistency is the thing it offers, and consistency is usually what breaks first in multi-area analysis.

Take the municipal file when you are working inside one municipality, you need new construction promptly, or block-level positional accuracy matters more than anything else.

Take both when you want to know how much the choice costs you. The comparison is often more informative than either file.

What the comparison shows

On any download page with more than one source available:

The map. Toggle between sources and the difference is immediate — addresses present in one and not the other, and positions that disagree, both show up as visible pattern rather than a number you have to interpret.

The summary alongside it:

Addresses in one source only. A municipal file missing addresses the register has usually means an annexed area or a lagging municipal extract. The register missing addresses the municipality has usually means new construction.

Positional disagreement. How far matched records sit apart. Small and uniform means a datum or rounding difference. Large and clustered means one source is placing records at a representative point.

Field coverage. Which fields each source populates, and how completely. This is where the postal code and unit gaps become visible.

None of this ships inside the download. It is on the page so the choice is informed before you take a multi-million row file, not after.

The obvious combination

Many people want municipal positions with register postal codes. That is a reasonable thing to want and you can build it yourself:

SELECT
  m.address_id,
  m.civic_number,
  m.street_name,
  m.geom          AS municipal_position,
  n.postal_code,
  n.unit_count
FROM municipal_addresses m
LEFT JOIN national_addresses n
  ON  n.civic_number = m.civic_number
  AND n.street_name  = m.street_name
  AND ST_DWithin(n.geom::geography, m.geom::geography, 50);

The 50 m tolerance is doing real work. Matching on address components alone produces false joins where a street name repeats; matching on position alone fails where the two sources place a record differently. Together they are reliable enough for most purposes.

We do not ship this join as a product. Doing it properly requires decisions about which source wins each field, and those decisions depend on your work rather than ours.

A caution about counting

Do not add the two files together. They overlap almost entirely, and the union is not a bigger dataset — it is the same addresses twice with different identifiers.

If you need one file covering a municipality, pick one source. If you need coverage beyond the municipal boundary, that is the national register by definition.