---
configs:
- config_name: default
  data_files: "data.tsv"
  names: ["kind", "sound"]
size_categories:
- n<1K
---

# [doc] formats - tsv - 3

This dataset contains one tsv file at the root:

- [data.tsv](./data.tsv)

```tsv
dog	woof
cat	meow
pokemon	pika
human	hello
```

We define the config name in the YAML config, the file's exact location, and the columns' name. As we provide the `names` option, but not the `header` one, the first row in the file is considered a row of values, not a row of column names. The delimiter is set to `"\t"` (tabulation) due to the file's extension. The reference for the options is the [documentation of pandas.read_csv()](https://pandas.pydata.org/docs/reference/api/pandas.read_csv.html).

```yaml
---
configs:
- config_name: default
  data_files: "data.tsv"
  names: ["kind", "sound"]
size_categories:
- n<1K
---
```