CLI Quickstart

This walkthrough checks a small customer dataset’s contract, reads two records, and exports its schema and CSV parsing settings. You can run every step in a terminal without writing application code.

You need a .NET 8, 9 or 10 SDK to install the CLI with the commands below. If you prefer an executable that includes .NET, follow the standalone installation instructions and then continue at the sample download step.

Install the CLI

dotnet tool install --global Packata-cli
packata --version

The first command installs packata as a global .NET tool. The second prints the installed version, confirming that your terminal can find it. If it is already installed, use dotnet tool update --global Packata-cli instead. If the command cannot be found, follow the installer’s PATH instructions for your shell.

Download the sample

Save datapackage.json and customers.csv in the same new directory, then open your terminal there. In a repository checkout, the originals are in docs/samples/quickstart.

datapackage.json is the contract: it describes a dataset named customers, with an integer id and a string name, and points to customers.csv. The CSV holds the actual records for Alice and Bob. Keep both files together so the relative CSV path can be resolved.

Check the contract

Run validation before reading the data:

packata contract validate datapackage.json

For this sample, the command prints Contract is valid. and exits with code 0. It checks the descriptor against its standard. It does not open the CSV or prove that every record matches the schema.

See what the contract describes

Inspect the customer dataset:

packata contract inspect datapackage.json --asset customers

--asset customers selects the dataset by its identifier. The result shows its id and name fields and the endpoint named customers-endpoint, which identifies where its data is stored. This step still reads only the contract.

Read two records

Preview the data referenced by that asset:

packata data preview datapackage.json --asset customers --limit 2 --output-format json

--limit 2 reads at most two records. --output-format json makes the result easy to inspect or pass to another program. You should see:

{"schemaVersion":1,"columns":["id","name"],"rows":[[1,"Alice"],[2,"Bob"]]}

Whitespace can differ. Unlike validation and inspection, preview opens the CSV. None of these commands changes either sample file.

For ODCS enum labels and DPv2 category labels, add --decode-categories to data preview. The category walkthrough provides complete descriptors for both standards, CSV data, and output with and without decoding. It also explains native PocketCsvReader matching and why labeled columns return strings, including integer fallback codes.

If you receive the descriptor and CSV together in dataset.zip, you can also run packata data preview dataset.zip --asset customers --limit 2 --output-format json without extracting it. See the zipped package example for the archive layout and resource paths.

Export a record schema

To give another tool a JSON Schema for customer records, run:

packata schema export datapackage.json --output-format json-schema --asset customers

The generated schema describes one record with an integer id and a string name. Export prints the schema; it does not validate the CSV against it. The contract guide shows how to save it to a file.

Export the CSV parsing settings

A tool reading the CSV also needs to know its delimiter and header settings. Export those from the endpoint you found during inspection:

packata dialect export datapackage.json --endpoint customers-endpoint

The resulting Table Dialect JSON descriptor specifies a comma delimiter, a header and the sample’s newline setting. It exports settings already in the contract; it does not infer them by sampling the CSV. The dialect guide explains how to save and interpret the result.

Continue with your own task

Use the data guide to infer a schema from records, or the contract guide to review a diff and assess compatibility between versions. The CLI reference lists all commands and options.

If you want to embed these operations in an application, follow the separate .NET Quickstart.