On November 30, 2026, Microsoft’s Export to Data Lake add-in for Dynamics 365 Finance and Operations stops exporting new or changed data, and Microsoft has said it will grant no extensions. Files already in your data lake stay where they are, so reports that read them keep opening, but their numbers stop moving. Microsoft’s recommended replacement is Link to Microsoft Fabric with low-latency sync. This guide covers what stops, what changes in the data your reports read, and a cutover plan for the weeks that remain.
Export to Data Lake: the dates that matter
- November 1, 2024: Microsoft retired Export to Data Lake, replacing it with Azure Synapse Link for Dataverse.
- October 15, 2026: the Delta Lake option in Azure Synapse Link is no longer available to new customers.
- November 30, 2026: Export to Data Lake stops permanently, with no extensions.
- December 2027: existing Synapse Link Delta Lake users must have moved to CSV output or Link to Fabric.
What stops when Export to Data Lake ends, and what keeps running
The deadline applies only to the Export to Data Lake add-in installed through Lifecycle Services (LCS), which uses change data capture to write CSV files into a data lake you manage. Link to Microsoft Fabric and Azure Synapse Link are not affected, and existing Fabric links keep running, including links not yet on low-latency sync.
After November 30 the add-in exports nothing new, and the files it already wrote stay in your lake. That is the risk: Power BI models, Synapse serverless views, and Azure Data Factory pipelines that read those files keep running without errors while serving data frozen at the cutoff. To spot it, compare the latest DataLakeModified_DateTime in a high-volume table, such as GeneralJournalAccountEntry, with today’s date.
One practical warning: in September 2026 Microsoft removed its Export to Data Lake documentation from Microsoft Learn, including the folder layout and the change feed format. Before you migrate, write down how your own lake is organized and which tables and columns each report uses, because Microsoft no longer publishes that reference on Learn.
Link to Fabric or Azure Synapse Link
Microsoft recommends Link to Microsoft Fabric with low-latency sync for most organizations, and Azure Synapse Link for teams that must keep exporting data to their own Azure storage account.
| Link to Microsoft Fabric | Azure Synapse Link | |
|---|---|---|
| Where the data lives | A Delta Parquet copy in Dataverse-managed storage, available in your Fabric lakehouse through shortcuts | Your own Azure storage account |
| What you pay for | Dataverse database capacity, plus Fabric capacity | Your Azure storage, compute, and integration tools |
| Delta tables | Yes | The Delta Lake option is closed to new customers from October 15, 2026; incremental CSV output remains |
| How fresh the data is | Microsoft’s Fabric blog reports under 15 minutes in most observed cases with low-latency sync; Learn still says incremental updates typically appear within an hour | Set by the incremental update interval you choose |
| Best for | Most teams, and anyone reporting in Power BI on Fabric | Teams that must keep their own data lake pipelines |
For teams that feed their own warehouse, Microsoft describes Synapse Link’s incremental CSV profile as its equivalent of the old Export to Data Lake change feeds.
What changes in the data your reports read
Microsoft’s transition guide says the new services keep the same data shapes and that existing pipelines can stay as they are. Microsoft’s own FAQ and table documentation list changes that say otherwise, so plan for these:
| Export to Data Lake | Link to Fabric |
|---|---|
| CSV files in folders by application area, module, and table type, with CDM metadata | Delta tables in a Fabric lakehouse, with a SQL analytics endpoint |
| RecID as an integer | Id holds the RecID converted to a GUID; an F&O field named ID becomes FnO_Id |
| Deleted rows removed at once | Soft deletes flagged with IsDelete, purged from Delta tables after 28 days |
| DataLakeModified_DateTime | SinkModifiedOn and SinkCreatedOn; sysrowversion works as a watermark |
| Change feed folder with the full change history | An optional Change Data Feed that keeps about 24 hours of changes |
| Time zone (TZID) and binary fields | Removed |
| Long text fields | Kept, but cut at 2,000 characters |
| Fields named with SQL reserved words | Renamed with a trailing character, for example Level becomes Level_ |
| Mixed-case table and column names | Lowercase names in the SQL endpoint, with case-sensitive collation |
Azure Synapse Link applies the same column changes (Id, FnO_Id, IsDelete, the Sink columns, the removed time zone and binary fields, and the 2,000-character cut) but writes to your own storage, and its incremental CSV profile, not Change Data Feed, is Microsoft’s stated equivalent of the change feed folder.
Two more details: with low-latency sync, empty strings stay empty strings instead of becoming NULL as they did on the earlier engine, and enum fields arrive as numeric codes whose labels sit in the GlobalOptionsetMetadata table, which is exported automatically.
Five changes that quietly move finance numbers
These raise no errors. They change totals, so build them into your reconciliation:
- Soft-deleted rows. Filter out rows where IsDelete is set, or deleted journal lines stay in your totals for up to 28 days.
- Master company data sharing. Shared tables are exported for all companies, so filter out the companies your old export did not include.
- Exchange rates. On builds older than Microsoft’s fix (the 10.0.38 to 10.0.40 platform updates), valid-time-state tables such as ExchangeRate exported only the currently valid rows, and a full refresh is needed after updating. Builds that meet the low-latency minimum include the fix, but confirm historical rates are present before you reconcile currency translation.
- Derived tables. A derived table such as DirPerson does not include its base table’s fields, so select the base table, DirPartyTable, as well.
- Daily kernel tables. 19 kernel tables, including DATAAREA and USERINFO, refresh once a day, so a company or user added today can be missing from joins until the next refresh.
Also expect string filters to become case-sensitive in the SQL endpoint: dataareaid = ‘usmf’ matches, and ‘USMF’ does not.
Data entities and BYOD
Export to Data Lake and BYOD often fed reports from data entities rather than tables. Entities can come across too: once enabled as a virtual entity with row version change tracking, each appears as a Dataverse table with an mserp_ prefix, which Link to Fabric can select. Many entities fail Microsoft’s validation rules, and entities built for data migration have no Track changes option at all. Check yours in System administration > Setup > Row version change tracking > Data entity row version change tracking validation report. Where an entity fails, select its underlying tables and rebuild the entity shape, for example with Microsoft’s FastTrack EntityUtil. On mserp_ entities, enum fields hold Dataverse option values rather than the original F&O enum values, so reports that filter on enum codes need translating, for example with Microsoft’s FastTrack enum translation.
BYOD users have more time, since no retirement date is set, but the same table and entity changes apply when they move.
Before you start: the prerequisites
In Dynamics 365 Finance and Operations
- A Tier-2 sandbox or higher, or an environment provisioned from an ERP-based template.
- Power Platform integration, required for every F&O environment since May 1, 2025.
- The Sql row version change tracking configuration key (Finance 10.0.34 or later), turned on in License configuration in maintenance mode and tested in a sandbox first, and Allow Row Version Change Tracking set on custom tables.
- For low-latency sync, at least 10.0.47 (platform 7.0.7858.157), 10.0.48 (platform 7.0.7996.100), or 10.0.49 (platform 7.0.8199.19). Microsoft lists minimum builds only for these three versions; if you are below the minimum for yours, apply its latest quality update first.
In Microsoft Fabric and Power Platform
- The System Administrator role in the Power Platform environment, and admin rights on the Fabric workspace.
- Fabric capacity in the same Azure geography as the Dataverse environment.
- Three Fabric tenant settings turned on: “Users can create Fabric items,” “Create workspaces,” and “Users can access data stored in OneLake with apps external to Fabric.”
- Fewer than 2,000 active Dataverse tables in the environment (above that, Microsoft says Link to Fabric can fail), and a Fabric workspace whose inbound access allows all networks.
- Dataverse database capacity for the copy. Microsoft’s example: syncing 500 GB could add about 100 GB. An F&O license adds 10 GB of Dataverse database capacity, plus an amount per user license.
A cutover plan for the remaining weeks
- Inventory what reads the lake. List every Power BI model, view, and pipeline that reads Export to Data Lake folders, with the tables and columns each one uses.
- Check versions and capacity against the prerequisites above.
- Create the Fabric link and add F&O tables. In Power Apps, go to Link data, then New link, then Link data via Fabric. F&O tables are not selected by default, so select them in the wizard’s Select Tables step or later in Manage tables. Run the setup off-peak. If the environment already has a Fabric link, look for the Low-latency mode flag on the Link data page: a link without it stays on the earlier engine, and moving it means unlinking and relinking, which deletes the lakehouse and re-syncs every table.
- Start with the largest tables. Microsoft gives no fixed initial sync time. It quotes about 1 million records per hour per table for low-latency sync in lab tests, so at that rate a table with 200 million rows takes about 8 days. Tables show as Unidentified in the lakehouse until their first sync completes.
- Run both and reconcile. Export to Data Lake keeps running until November 30, so compare row counts and month-end totals, such as the trial balance and the receivables and payables subledgers, across the October month-end close, the last one before November 30.
- Repoint reports. Rebuild semantic models on the lakehouse, ideally in Direct Lake mode, which reads the Delta tables directly instead of importing a copy. If you built Synapse serverless views over the old data, don’t repoint them at the new tables, because Microsoft says serverless SQL pools can’t read low-latency Delta tables in all scenarios; Microsoft’s FastTrack Fabric Link transition tool can recreate them in a Fabric warehouse with the old shapes.
- Archive and uninstall. Keep the frozen lake for audit, since the old change history does not move to Fabric. Then uninstall the add-in in LCS under Environment Add-ins > Export to Data lake > Uninstall; it needs no downtime.
If you miss the deadline
A late start loses no live F&O data, because the initial sync in Link to Fabric copies each table’s current contents from F&O. What stays behind is the old change history in the frozen lake, and reports stay frozen until the initial sync and the report rebuild finish, which can take days for large tables. Until you cut over, label every report that reads the old files as frozen at November 30, so no one acts on stale numbers.
Where Metrixs fits
Link to Fabric gives you tables, not reports. Metrixs runs in your own Microsoft Fabric environment on Link to Fabric or Azure Synapse Link and adds its data model and 150+ pre-built Power BI reports, with 1,200+ metrics and 250+ KPIs. On the Standard plan, reports use Direct Lake and refresh every 15 to 30 minutes, depending on data volume, the number of reports, and Fabric capacity. See Microsoft Fabric for D365 F&O for how it fits together, compare the options in Power BI vs other D365 reporting tools, or request a demo and see if you qualify for a free 6-month trial.
FAQs
1. When does Export to Data Lake stop?
November 30, 2026. After that it exports no new or changed data, files already in your lake stay where they are, and Microsoft grants no extensions.
2. Does the deadline affect Link to Fabric or Azure Synapse Link?
No. It applies only to the Export to Data Lake add-in installed through Lifecycle Services. Link to Fabric and Azure Synapse Link keep running, although the Delta Lake option in Azure Synapse Link is closed to new customers from October 15, 2026, and existing users must move off it by December 2027.
3. Will our existing Power BI reports keep working?
They keep running against the files already in your lake, but the data stops at November 30. Repoint them to Link to Fabric and plan for the schema changes: GUID Id fields, soft deletes, removed time zone and binary fields, and lowercase names in the SQL endpoint.
4. Is BYOD being retired too?
Microsoft has set no retirement date for bring your own database (BYOD), but it recommends moving to Link to Fabric or Azure Synapse Link.
5. How long does the initial sync take?
Microsoft gives no fixed time. It depends on table size and the number of tables. Microsoft quotes about 1 million records per hour per table for low-latency sync in lab tests, so at that rate a table with 200 million rows takes about 8 days.
Sources
All pages below were checked on October 14, 2026.
- Removed or deprecated platform features: Export to Azure Data Lake
- FAQ: transitioning from legacy data integration services
- Transition from legacy data integration services to Fabric link and Azure Synapse Link
- Choose finance and operations data in Azure Synapse Link for Dataverse
- Configure your environment and link to Microsoft Fabric
- Microsoft Fabric link FAQ
- Azure Synapse Link Delta Lake option deprecation
- Azure Synapse Link incremental updates
- Row version change tracking for tables and data entities
- Enable Power Platform integration
- Microsoft Power Platform blog: low-latency sync, June 9, 2026
- Microsoft Fabric Updates blog: what’s new for Dataverse Fabric link, July 2026
- Microsoft FastTrack: Fabric Link and the SQL analytics endpoint (GitHub)
- Microsoft FastTrack: data integration with Dataverse link (GitHub)
- Microsoft Learn source repository: deletion of the Export to Data Lake articles, September 8, 2026 (GitHub)
Senior Business Intelligence Architect with 22 years of experience designing enterprise analytics and data platforms. Focus areas include Power BI, real-time analytics, and large-scale BI architecture across the Microsoft data stack.
