skipprd schema
skipprd schema prints the columns and Arrow types of the Parquet files in a pipeline's local output folder, one name: type line per column, sorted by name. Use it to check what actually landed on disk when you write Parquet locally, for example with the Local file destination. To see the schema discover inferred and saved for any pipeline, use skipprd metadata show instead; it works before your first sync and for every destination.
Usage
skipprd schema --pipeline <name> [--config <path>] [--log [LEVEL]]There is no Python equivalent. In Python, run session.query(...) or session.df() and inspect the returned Arrow table's schema.
Options
| Flag | Default | Description |
|---|---|---|
-p, --pipeline <NAME> | Required | Pipeline whose output to read. Unlike discover and sync, schema does not fall back to PIPELINE_NAME. |
--config <PATH> | ./skippr.yml | Config file to read. |
--log [LEVEL] | Off | Print logs to stderr, including the folder being read. |
--wal-storage <MODE> | disk | Must not be clustered. |
See CLI overview for the other global flags.
Where it reads
schema reads every .parquet file under <data dir>/<workspace>_<pipeline>/output. The data dir is the pipeline's data_dir in skippr.yml, or the DATA_DIR environment variable, default ./data. For workspace demo and pipeline rides, that is ./data/demo_rides/output.
To make the Local file destination write there, set its output_dir to that folder.
Examples
Print the schema of local Parquet output
skipprd schema --pipeline ridesduration_s: Int64
member: Boolean
ride_id: Int64
station: Utf8Show the folder being read
skipprd schema --pipeline rides --logThe log line Querying data dir: ./data/demo_rides/output shows exactly which folder schema scanned.
See the discovered schema instead
skipprd metadata show --pipeline ridesTroubleshooting
| Symptom | Cause | Fix |
|---|---|---|
Panics with listing table: Internal("No schema provided.") and exits 101 | There are no Parquet files in the output folder. This is normal for destinations that write elsewhere (warehouses, S3, Iceberg). | Use skipprd metadata show to see the discovered schema, or query the data with skipprd query. |
error: the following required arguments were not provided, followed by --pipeline <PIPELINE> | schema needs --pipeline on the command line. | Pass --pipeline <name>. |
WAL_STORAGE=clustered does not support schema | schema does not run in clustered mode. | Run it without --wal-storage clustered. |