denis256 notes and projects

ingestr v1.0.45, published on June 24, 2026, is a focused release for destination naming and Braze exports. The main user visible change is support for three component destination names, so warehouse targets can carry catalog, schema, and table parts without being flattened into the wrong shape. It is not marked as a prerelease.

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

Three Component Destination Names

The important naming change in this release is support for three component table names. In practical terms, ingestr can now handle destination names that need catalog, schema, and table parts instead of forcing everything through a shorter shape.

This matters for users who run ingestr into systems where catalog is not cosmetic metadata. Before this release, two level names could be confused with catalog qualified values in the surrounding naming code. v1.0.45 fixes that distinction: two level destination names are not treated as catalog qualified, while three component names can represent catalog.schema.table.

That is small but sharp. A data load tool should not guess its way through warehouse naming. If the configured target says three parts, the loader should preserve three parts. If it says two, the loader should not promote it into a catalog form by accident.

Multi Table Names Handle Source Schema Dots

Another user visible change is how multi table destination names deal with dots in source schemas. The release notes say source schema dots are flattened in multi table destination names. That reduces ambiguity when a source system uses dotted schema names and the destination naming logic is already using dots as separators.

Without this kind of cleanup, one source schema can look like multiple namespace levels after it lands in the destination. Operators then get annoying table names, or worse, table names that look valid but point at the wrong logical place. Ingest jobs are boring only when naming is boring. This release makes that part more deterministic.

Braze User Data Exports All Segments by Default

Braze also gets a behavior change. For user_data, ingestr now exports all segments when no segment is specified.

This is a good default for batch style extraction because an empty segment filter now means full coverage instead of a too narrow run. It also means teams should check existing Braze jobs where the absence of a segment was accidental. The release notes do not describe a breaking change, but this default can still change the amount of data moved.

That last point matters for cost and runtime. If a job was running without segment configuration, compare expected row counts and destination volume after upgrading. The change is useful, but it makes implicit config more visible.

Naming Code Cleanup

The remaining changes are mostly internal naming cleanup. The maintainers describe catalog aware naming everywhere and removal of a multi table gating workaround.

This is not a feature to call from the CLI. It is still relevant for operators because the code path that decides final destination names is now less split across special cases. When naming bugs happen, they tend to show up as duplicated tables, missed tables, or jobs that write to a surprising namespace. Reducing workaround logic is a practical maintenance win.

The release also updates licenses.lock.yml for merged streaming connector dependencies. That reads like dependency bookkeeping, not a runtime change. It is worth noting only because license lock files often affect build and packaging checks.

Where to get it

• Release page: ingestr v1.0.45
• Repository: project repository
• Tag: v1.0.45