> For the complete documentation index, see [llms.txt](https://knowledgebase.fabricdata.com/fabric-knowledge-base/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://knowledgebase.fabricdata.com/fabric-knowledge-base/learn-the-application/destinations/collections/collection-export.md).

# Collection Export

The collection export feature has been improved to address previous inconsistencies and performance issues. Previously, large collections often resulted in incomplete exports or timeouts, generating multiple support tickets. With the new implementation in the Record Service, exports are now faster, more reliable, and able to handle large datasets efficiently.

\
Related Tickets:

MSD: [MSD-8886](https://fabric.atlassian.net/browse/MSD-8886), [MSD-10048](https://fabric.atlassian.net/browse/MSD-10048), [MSD-8987](https://fabric.atlassian.net/browse/MSD-8987), [FABRIC-8475](https://fabric.atlassian.net/browse/FABRIC-8475)

Fabric: [FABRIC-7964](https://fabric.atlassian.net/browse/FABRIC-7964), [FABRIC-8316](https://fabric.atlassian.net/browse/FABRIC-8316), [FABRIC-8319](https://fabric.atlassian.net/browse/FABRIC-8319), [FABRIC-8475](https://fabric.atlassian.net/browse/FABRIC-8475), [FABRIC-8525](https://fabric.atlassian.net/browse/FABRIC-8525)

***

### Key Enhancements

1\. Faster and More Reliable Exports

* Large collections now export quickly and efficiently, with significantly reduced risk of timeouts.
* The export process uses smarter data handling to avoid unnecessary duplication and heavy memory usage.
* Users can now export even very large collections without performance issues.

\
2\. Complete Metadata

* Exports now include all linked records, including bidirectional linkages, ensuring that the data is comprehensive and accurate.
* Column counts and headers are calculated efficiently, so all expected information appears correctly in the exported file.

\
3\. Reduced Memory Usage

* Data is processed in manageable chunks rather than loading all records at once.
* The export file is streamed directly to the client, avoiding the memory-intensive Excel build used previously.
* These improvements reduce server load and improve response times.

\
4\. Better Compatibility and Structure

* Exports are now UTF-8 encoded with BOM, ensuring proper display in Excel and other tools.
* The new pipeline (Record Service + MetadataExportService) provides a clearer, more maintainable structure.

***

### Bottom Line

The new collection export path:

* Faster: Optimized data processing and SQL-based calculations.
* More reliable: Handles large datasets without timeouts.
* Complete: All linked records included, bidirectional links covered.
* Memory-efficient: Chunked processing and streamed download.

<br>

This implementation ensures a more efficient, reliable, and accurate export experience for all users.