# Resources

Resources represent data files that can be downloaded from a provider.

## Single file download

Download a single file from a URL.

→  Use file resources such as [](http-resources.md) for standard HTTP resources or [](sdmx-resources.md) when the file is served by an SDMX 2.1 web service (adds SDMX error handling and header normalization).

## Couple of files download

Download several files that form a logical unit (e.g., a data file and its structure file).
If any file fails, none are written to the target directory — the group is all-or-nothing.

→ Use [](resource-groups.md)

## Paginated download

Download resources from an HTTP API paginated with `?page=N` where the total number of pages is unknown.
The group enumerates pages starting from `start_page_num` and stops when the resource raises [](#PageOutOfBounds).

There is **no split**: each page corresponds to exactly one resource, and failures are not retried as smaller sub-pages.

→ Use [](page-resource-groups.md)

## Partitioned download

Download a large dataset by splitting it into partitions (e.g., filtered by time period, by region).
Unlike pagination, partitions can be **split dynamically**: when a resource fails (e.g., [](#ResponseTooLarge)), it can raise [](#EnqueueSubPartitions) to replace 1 resource with N smaller ones.

→ Use [](partition-groups.md)

```{toctree}
:caption: Contents:
:hidden:

base-classes
http-resources
sdmx-resources
resource-groups
page-resource-groups
partition-groups
resource-id
```