Data Extract API: keeping a downstream copy in sync

The Data Extract API (developers.rtafleet.com) returns raw rows from Fleet360 tables for use in data warehouses and integrations. This article covers how to keep a downstream copy in sync when you pull data incrementally.

How an incremental pull works

Every row in an extractable table carries an etag - a value that increments whenever any column in the row changes. When you call the extract endpoint with an etag parameter, the API returns only rows that were created or changed since that marker, sorted by etag, and hands you back a nextEtag for your next call.

The etag bookmark, exactly

  • The etag parameter is inclusive: a request returns all rows with etag >= the value you supply, in ascending order. Reusing an older etag re-pulls those rows.

  • nextEtag is the first row not included in the returned page - each page resumes exactly where the last one stopped. Save it and pass it as the next request’s etag.

  • When a page returns no nextEtag you have reached the end. The correct bookmark for your next refresh is the last returned row’s etag + 1. Example: if the final page ended at etag 101944492, the next refresh starts with etag=101944493.

  • An empty refresh (nothing changed since your marker) returns count: 0 and no nextEtag. Keep your saved etag and retry on your next schedule - do not reset it to 0.

  • Every row in the response is the complete record - all columns, including nulls - not just the fields that changed. Replace the whole row in your copy when it comes back.

What an incremental pull does not return

Incremental pulls return rows that exist and were added or changed since your marker. There is no separate deletion notice: when a row is removed from the Fleet360 table it simply never appears in a pull again. A downstream copy that is only fed by incremental pulls keeps that row forever - this is how a copy ends up with extra records that no longer exist in Fleet360.

Archived is not the same as deleted

An archived asset is still a real row - it is returned by pulls and shows in Fleet360 under the archive view, just flagged as archived. A deleted record no longer exists at all. If your downstream count is higher than Fleet360’s, do not assume the difference is archived assets: compare against a current full pull first. Rows your copy has that Fleet360 does not return at all are deleted records; rows that come back with an archive flag are archives and need different handling.

Reconcile with a periodic full pull

  1. Run a full pull: call the extract endpoint with etag=0 and page through with nextEtag until it is no longer returned.

  2. Remove any row from your copy whose key is not in the current result.

  3. Repeat on a schedule that fits how exact your copy needs to be (weekly or monthly is typical), or whenever counts drift.

If your counts do not match, check whether the extra rows in your copy are (a) real Fleet360 rows flagged archived - handle them per your integration’s archive rules, or (b) rows Fleet360 does not return at all - deleted records, removed by the next full reconciliation.