denis256 notes and projects

ingestr v1.0.38 was published on June 19, 2026, with two practical source additions: SQL Server Change Tracking and SharePoint. The release also tightens sports data sources, with schema inference work and safer interval handling across api-football and football-data. It is a regular release, not a prerelease.

The full release notes and downloads are on the GitHub release page.

SQL Server Change Tracking and SharePoint arrive

The headline user change is source coverage. The changelog lists SQL Server Change Tracking in PR #808 and a SharePoint source in PR #817. For teams using ingestr as a small moving part in data ingestion, that means less glue code around common Microsoft systems.

SQL Server Change Tracking matters because it gives a database native way to capture row changes without treating the whole table as fresh input. The release notes do not spell out every configuration detail, so operators should still read the source docs before wiring it into scheduled jobs. The useful part is simpler: this source now exists in the project line, so pipelines that already standardize on ingestr have a more direct option.

The SharePoint source targets a different pain. Many teams still receive operational data through SharePoint files or lists. Pulling that into the same ingestion tool as database and API sources reduces another separate script.

Sports sources get schema inference

The release also spends real time on sports related connectors. api-football and football-data both move toward schema inference, and BallDontLie gets a refactor with schema inference, a replace strategy, and a scheme rename to balldontlie.

That is mostly plumbing, but it is user facing plumbing. Schema inference reduces the amount of fixed schema work maintainers must keep in sync with APIs that can be uneven. The changelog also notes a rename from api-football to apifootball, and from football-data to footballdata. Existing configs that refer to the old scheme names deserve a check during upgrade.

Interval handling replaces date parameters

Several commits clean up how football data intervals are passed around. The release drops date_from and date_to URI params in favor of interval. It also changes fixture derived api-football tables so merge behavior honors the interval.

That is the kind of change that can bite scheduled jobs if it is ignored. If a pipeline builds source URIs directly, check for those older params. The target behavior is clearer: one interval drives the read, and merge only applies where the table can be loaded incrementally.

Safer keys and rate limits

Two smaller items are worth noticing because they reduce noisy failures. The release protects match_events synthetic keys from collisions with API fields. It also tightens cancellation behavior after review.

On the rate side, api-football lowers its burst setting to stay under the free tier cap, and docs lose a stale plan quota line. That is not a flashy feature, but it is exactly the sort of correction that keeps a connector from behaving fine in tests and then failing under a real quota.

Upgrade notes

Read this release as a source expansion plus connector cleanup release. The main migration risk is scheme naming and date parameter handling around the sports sources. Search configs for api-football, football-data, date_from, and date_to, then test the new apifootball, footballdata, and interval shape before changing production schedules.

Also pay attention to primary keys if any downstream process depends on flattened helper output. The changelog includes a fix for stale primary keys after removing flatten helpers, which suggests this area changed enough to deserve a basic data quality check after upgrade.

Where to get it

• Release page: ingestr v1.0.38 release
• Repository: project repo
• Tag: v1.0.38