# SDMX resources

[](#SdmxResource) extends [](#HttpResource) to handle [SDMX 2.1](https://sdmx.org/) web service responses. It adds SDMX-specific error handling and XML header normalization on top of the standard HTTP resource capabilities.

## Constructor parameters

In addition to all parameters inherited from [](#HttpResource) and [](#FileResource), [](#SdmxResource) accepts:

- `normalize_sdmx_file_header` (`bool | None`, default: `True`) — whether to normalize the SDMX file header after download (see below).
- `sdmx_parser_factory` (`SdmxParserFactory | None`) — a factory for creating SDMX parsers. Defaults to `SdmxParserFactory()`.

## SDMX error handling

SDMX web services can return errors in the XML response body that differ from the HTTP status code. The method `raise_for_sdmx_error(error)` inspects the response content and raises specific exceptions:

- [](#NoResultFound) — when the SDMX error code is `NO_RESULT_FOUND` (the query returned no data).
- [](#ResponseTooLarge) — when the SDMX error code is `RESPONSE_TOO_LARGE` (the requested dataset exceeds the server's size limit).
- [](#SdmxResourceError) — for any other SDMX error code.

These exceptions can be caught by a [](#BasePageResourceGroup) or [](#BasePartitionGroup) to trigger a split or stop pagination.

### Example: handling response-too-large in a partition group

```python
from dbnomics_toolbox.fetcher_utils.processors.errors import EnqueueSubPartitions
from dbnomics_toolbox.fetcher_utils.resources.errors import ResponseTooLarge


class MyPartitionGroup(BasePartitionGroup):
    def _download_partition(self, partition, *, partition_depth, partition_num):
        try:
            super()._download_partition(
                partition,
                partition_depth=partition_depth,
                partition_num=partition_num,
            )
        except ResponseTooLarge:
            # Split the partition into smaller ones
            raise EnqueueSubPartitions(partition.split()) from None
```

## SDMX header normalization

When `normalize_sdmx_file_header=True` (default), the `_post_process_target_file` method calls [](#normalize_sdmx_file_header) after the standard MIME type validation and file reformatting.

This normalization ensures that SDMX XML headers are consistent across different versions of the same file, reducing noise in version control diffs (e.g., the `Prepared` timestamp is normalized to a fixed date).

## Download flow

1. [](#HttpResource) fetches the URL and writes the response to the target file.
2. [](#FileResource) validates the MIME type and reformats the file (JSON/XML).
3. [](#SdmxResource) normalizes the SDMX file header.

If any of these steps fails, the resource is considered failed.