Learn
UWI formats: US vs. Canada
Every well needs exactly one identity. The US and Canada solved that problem in opposite ways — a sequential registry number versus an identifier that encodes the well's location — and mixing up their rules corrupts databases quietly.
By WellSchema Engineering ·
A unique well identifier (UWI) is the primary key of the oilfield: the one string that regulators, partners, data vendors and your own databases agree refers to the same hole in the ground. North America has two incompatible conventions, and most well-header data problems trace back to treating one like the other.
The US: the API well number
The US uses the API well number (today formally the US Well Number, stewarded by the PPDM Association after the API retired the standard). It is a sequential registry number: it tells you where to look a well up, not where the well is. The base number is 10 digits, extended to 12 or 14 for sidetracks and well events:
| Digits | Field | Example (42-165-01234-01-02) |
|---|---|---|
| 1–2 | State code | 42 = Texas |
| 3–5 | County or parish code | 165 = Gaines County |
| 6–10 | Unique well serial, assigned by the state regulator | 01234 |
| 11–12 (optional) | Directional sidetrack code — 00 is the original hole | 01 = first sidetrack |
| 13–14 (optional) | Event sequence code (recompletions, operations) | 02 |
Punctuation is not part of the standard — 4216501234, 42-165-01234 and 42-165-01234-01-02 can all refer to the same wellbore at different levels of detail. Databases should store the digits and treat dashes as display formatting.
Canada: a location-based UWI
Canada took the opposite approach: the 16-character Canadian UWI (a CAPP/PPDM standard) encodes the well's legal survey location. For wells on the Dominion Land Survey (Alberta, Saskatchewan, Manitoba and parts of BC), the familiar display form is:
- Position 1 — survey system: 1 = DLS, 2 = NTS (British Columbia), 3 = federal permit areas, 4 = geodetic coordinates.
- Location exception + legal location: the LSD / section / township / range / meridian chain that puts the well on the map — the same string a landman would quote.
- Event sequence (last position): distinguishes drilling and completion events at the same location — the counterpart of the US 13th–14th digits.
Because the UWI encodes location, two wells drilled a section apart get visibly different identifiers — and a typo usually produces a location that does not exist, which makes Canadian UWIs partially self-validating.
Side by side
| US API / US Well Number | Canadian UWI | |
|---|---|---|
| Length | 10, 12 or 14 digits | 16 characters |
| Encodes | Registry: state, county, serial | Location: survey system + legal location |
| Assigned by | State regulator | Provincial regulator (AER, BCER, …) |
| Sidetracks / events | Digits 11–12 and 13–14 | Event sequence character(s) |
| Self-checking? | No — any 10 digits look plausible | Partially — invalid locations stand out |
The pitfalls that actually corrupt data
- Storing well numbers as integers. State codes and serials keep leading zeros; an integer column silently turns 05-001-00123 into 500100123. UWIs are text, always.
- Mixing display and storage formats. Decide once whether dashes/slashes are stored or applied at display time, and validate on entry — a format regex catches most transposition errors at the door.
- Ignoring the event/sidetrack suffix. The 10-digit API number identifies a surface well; the sidetrack digits identify the wellbore. Joining datasets at the wrong level double-counts production or loses laterals.
- Assuming one country's rules. A validator that accepts "anything numeric" will happily accept a Canadian UWI with its letters stripped — and now the record matches nothing.
The fix is unglamorous: validate the format on entry against the convention the well actually belongs to, store the canonical string, and format for display separately. It is a few lines of code that saves years of reconciliation.