# Parquet | DRE

> The parquet format.

You're reading the docs for DRE 0.4 (v0.4.4). [This page for 0.5](https://getdre.com/docs/plugin-parquet/) · [New in 0.5](https://getdre.com/docs/new-in-0.5/)

# Parquet

Parquet: the `parquet` format, from the `parquet` package. Declare the package once in `dependencies.yml`; DRE installs it on demand:

dependencies.yml

```yaml
plugins:
  - parquet
```

### Options

None.

## Types

Arrow types are preserved: a decimal stays a decimal at its precision, a timestamp keeps its time zone.

## Keep types in an analytics export

Use Parquet when the receiver needs typed data rather than display formatting. Starting from the [tutorial project](https://github.com/get-dre/dre/blob/v0.4.4/examples/tutorial/), add `parquet` to `dependencies.yml` and change:

```yaml
output: {format: parquet}
```

One query makes one `.parquet` file; several queries make separate files. Preview the output and read it back with DuckDB to verify types:

```sql
select * from read_parquet('path/to/the/generated/file.parquet')
```

Use the path printed by the run or `dre history <report> --latest --path`; run folder names change each time. The [monthly finance example](https://github.com/get-dre/dre/blob/v0.4.4/examples/monthly-finance/) supplies a realistic dataset to export by changing its format. Remove xlsx-only query `columns` and `style` settings when doing so. `invalid-output-option` identifies options belonging to another format; `undeclared-plugin` means `parquet` must be added to dependencies. See [output keys](https://getdre.com/docs/v0.4/reference-report/#output), [error codes](https://getdre.com/docs/v0.4/reference-error-codes/) and [format](https://getdre.com/docs/v0.4/glossary/#format).

[Previous  
Google Chat](https://getdre.com/docs/v0.4/plugin-google_chat/)[Next  
PostgreSQL](https://getdre.com/docs/v0.4/plugin-postgres/)