Folder tree, more formats, clearer errors
Browse your bucket as a folder tree. The Point at your data step is now an expanding tree, and you can type a prefix to jump straight to a folder.CSV and newline-delimited JSON. The wizard now accepts CSV (comma-separated, with a header row) and newline-delimited JSON alongside Parquet. A CSV with another delimiter is refused at setup, naming the header it read, instead of arriving as a single column.No duplicate pipelines. A folder that’s already connected is marked as ingested and can’t be selected again.Errors you can act on. Invalid input is refused at the field where you typed it, not several steps later, and access checks report exactly which permission is missing.Policies as commands. AWS role grants are presented as ready-to-copy commands instead of files to download.More mapping control. Resolve a table’s user ID by joining it to another table, skip a table from the mapping screen, and name the event for a folder that has no event-name column.Events sync as a daily full refresh. Event files from a date folder are now always read as a full refresh, so a re-run replaces the day instead of double-counting it.
AI-assisted schema mapping for file sources
AI-assisted mapping. Every file source can use Auxia’s schema-mapping agent, which reads your files, identifies the table type, and proposes a mapping with a confidence level for you to review.