
Reads Apache Parquet files from the bound directory and emits one Analytics.TabularDataset per file, optionally projecting a subset of columns. Use it to pull a columnar dataset into a tabular pipeline.
Load Parquet, serialise it, and forward the rows as JSON.
Replay a Parquet directory and log each dataset.
columns left empty reads every column, while a non-empty list projects only the named columns and an unknown name fails the read.repeat off each file is read once and the component idles; with it on the directory is re-read with check_interval seconds between passes, and a non-positive interval produces a tight continuous re-read loop.jthread reader (stop+join) + docs lint