CLI overview

skipprd is the Skipprd command-line tool. You use it to build a skippr.yml, discover the shape of your source data, sync it to a destination, and query the result. Every command reads the same skippr.yml, and the Python skippr.Session runs the same engine, so you can mix the CLI and Python freely.

Skippr, your data department

Speak to an expert. Upfront insights in days, not months.

Book an intro call →

20 minutes · No sales pitch

Check your install

skipprd --version

You see the installed version, for example:

skipprd 17.0.0

If the command is not found, follow Install.

Commands

Command	Use it to
skipprd connect	Add a source, destination, or schema sink to skippr.yml without hand-editing YAML.
skipprd discover	Read the source and infer the schema of each namespace. Never writes to the destination.
skipprd schema	Print the columns of Parquet files a pipeline wrote to its local data directory.
skipprd metadata	Show the schema discover saved, or apply a reviewed schema for one namespace.
skipprd sync	Ingest from the source, buffer through the write-ahead log (WAL), and land data in the destination.
skipprd query	Run SQL against your pipelines, interactively or as JSON for scripts.
skipprd df	Print every row of a pipeline or namespace as JSON.
skipprd serve	Expose your Iceberg tables over Iceberg REST and Flight SQL to tools such as Spark, Trino, and PyIceberg.
skipprd doctor	Check skippr.yml for mistakes before a long sync.
skipprd sql-help	List the SQL statements skipprd query supports, or export them to a file.
skipprd benchmark	Generate a synthetic JSON dataset for performance testing.

Run skipprd <command> --help to see every flag for a command.

Global flags

Global flags work with every command. You can put them before or after the command name: skipprd --config prod.yml sync and skipprd sync --config prod.yml are the same.

Flag	Default	Description
--config <PATH>	./skippr.yml	Config file to use. Same as the SKIPPR_CONFIG_FILE environment variable. Without it, Skipprd looks in the current directory for skippr.yml, then skippr.yaml. If the path you pass does not exist, Skipprd falls back to that search.
--log [LEVEL]	Off	Print logs to stderr. LEVEL is trace, debug, info, warn, or error; --log on its own means info. Logs are always written to logs/skipprd.<date>.log, with or without this flag.
--wal-storage <MODE>	disk	Where the WAL keeps buffered batches: disk, s3, or clustered. Same as WAL_STORAGE. See WAL and buffering.
--wal-s3-bucket <BUCKET>	State bucket	S3 bucket for WAL segments when --wal-storage s3. Same as SKIPPR_WAL_S3_BUCKET. A skippr.wal_s3_bucket value in skippr.yml takes precedence.
--store-type <TYPE>	sled	State store for offsets and run leases: sled (the local on-disk state store) or dynamodb. Same as SKIPPR_STORE_TYPE. A skippr.store.type value in skippr.yml takes precedence. See State store.
--store-name <NAME>	None	Table name for the dynamodb state store. Same as SKIPPR_STORE_NAME. Requires --store-type when used with connect.
--workspace <NAME>	None	Written to skippr.workspace by skipprd connect. Other commands ignore it and read skippr.yml.
--storage-mode <MODE>	None	local or s3. Written to skippr.skipprd_el_storage_mode by skipprd connect. Other commands ignore it.
--skippr-s3-bucket <BUCKET>	None	Written to skippr.skippr_s3_bucket by skipprd connect. Other commands ignore it.
--tenant <NAME>	None	Written to skippr.tenant by skipprd connect. Other commands ignore it.
-h, --help		Print help for skipprd or any command.
-V, --version		Print the version. Top level only.

Skipprd also loads a .env file that sits next to your skippr.yml, so ${NAME} references in the config can resolve from it.

Typical workflow

A pipeline goes from nothing to queryable data in five steps. The example uses the public bike-hire sample and keeps engine state on local disk.

1. Connect a source. Write the pipeline into skippr.yml:

   skipprd --workspace bikehire --storage-mode local connect data-source s3 \
     --pipeline bikehire \
     --name sample \
     --s3-bucket skippr-public-sample-data \
     --s3-prefix bike-hire

2. Check the config.

   skipprd doctor

3. Discover the schema. Skipprd reads the source and saves the inferred columns for each namespace:

   skipprd discover --pipeline bikehire --output text
   skipprd metadata show --pipeline bikehire

4. Sync. Run one pass and exit:

   skipprd sync --pipeline bikehire --once --output text

5. Query. Tables are named <pipeline>.<namespace>:

   skipprd query --plain --sql "SELECT count(*) FROM bikehire.rides"

The same workflow in Python:

Python

import skippr

session = skippr.Session(skippr.Config.discover().get_pipeline("bikehire"))
print(session.doctor())
session.discover()
session.sync(once=True)
print(session.query("SELECT count(*) FROM bikehire.rides"))

CLI

skipprd doctor
skipprd discover --pipeline bikehire
skipprd sync --pipeline bikehire --once
skipprd query --plain --sql "SELECT count(*) FROM bikehire.rides"

Output modes

Pick the output that suits who is reading it:

Command	Modes	Use for scripts
discover, sync	--output progress (default; a live progress display on a terminal), text, json	--output json prints one JSON event per line.
doctor	--output text (default), json	--output json
query	Interactive table view (default), --plain	--plain prints one JSON document: {"header": [...], "rows": [[...]]}.
df	JSON only	Same shape as query --plain.
metadata show, metadata apply	JSON only	Always JSON.

Exit codes

Code	Meaning
0	The command succeeded.
1	The command failed: invalid or unreadable config, a missing ${NAME} variable, a failed check, sync, discover, or query.
2	The command line was invalid, for example an unknown flag or a missing required flag.

Some failures are only described in the logs. If a command exits 1 without explaining why, run it again with --log.

Next steps

• Install
• skippr.yml reference
• How Skipprd works
• Python