Data Pump (expdp/impdp)
Lives on unchanged in Azure.
✓ certified by the mortician
In life
Data Pump is Oracle's server-side export and import framework: the expdp and impdp clients, the DBMS_DATAPUMP job API and the DBMS_METADATA API for extracting object definitions. Jobs run inside the database, write proprietary dump files to directory objects, can run in parallel on Enterprise Edition, can move data over a database link without files, and can transport whole tablespaces by copying data files.
Migrations care in two ways. For Oracle-to-Oracle moves to Oracle Database@Azure or Oracle on Azure virtual machines, Data Pump is the workhorse. For heterogeneous moves its dump files are useless to Azure SQL or PostgreSQL, but the metadata API is a reliable way to extract DDL for conversion tools, and existing Data Pump jobs and directory objects are part of the operational inventory.
Data Pump is supported and not deprecated; it is the replacement Oracle names for the original Export utility that the 26ai Upgrade Guide desupports.
Cause of departure
Not deprecated by Oracle. Listed here because Azure offers no clean equivalent.
Obituary
Data Pump, known familiarly as expdp/impdp, arrived before 11.2 and spent its life exporting and importing Oracle schemas and data. Its proprietary dump files carried Oracle estates with dependable self-possession.
It departs at Azure’s border: PostgreSQL Flexible Server accepts no Oracle source, while Azure Database Migration Service offers Oracle only for offline, preview migrations through SSMA to selected SQL targets. No Oracle online migration scenario is listed.
It is survived by Azure Database Migration Service, partly; Azure SQL Managed Instance, only by workaround; Ora2Pg, partly; Oracle Database@Azure, where it lives on unchanged; and SQL Server Migration Assistant for Oracle, partly. None reads its dump files outside Oracle.
Survived by
Oracle Database@Azure exact effort S
Oracle Database@Azure runs unmodified Oracle Database on Exadata infrastructure operated by OCI, so expdp and impdp work as on any Oracle system, including network-mode imports over a database link. Dump files are typically staged in OCI Object Storage or Azure Blob Storage mounted or accessed through the database, and CDB/PDB administration happens in the OCI console.
- Provision the Exadata VM cluster and target database from the Azure portal and OCI console.
- Stage dump files in object storage reachable from the database nodes, or use network mode with a link to the source.
- Run impdp with the usual parameters; no type mapping is involved.
Azure Database Migration Service partial effort M
Azure Database Migration Service lists Oracle only as an offline, preview source for Azure SQL Database, Managed Instance and SQL Server on Azure VMs, and the scenario is documented as running through SSMA. It offers no Oracle online (continuous sync) scenario and no Oracle to PostgreSQL scenario, and like every non-Oracle tool it does not consume Data Pump dump files.
- Use DMS with SSMA for offline Oracle to Azure SQL migrations where a managed, repeatable pipeline is wanted.
- Plan a downtime window; there is no Oracle online migration path in DMS.
- For PostgreSQL targets use Ora2Pg or the VS Code schema conversion feature plus Data Factory for data.
Complications
Ora2Pg partial effort M
Ora2Pg is the Data Pump equivalent for PostgreSQL targets: it connects to Oracle, exports schema objects by type and streams table data as COPY or INSERT files (or directly into PostgreSQL), with parallel jobs, table filters, WHERE clauses and type overrides. It cannot read dump files, and unlike Data Pump the export is a conversion, so type mapping decisions (NUMBER, DATE, ROWID) are made during export.
- Generate a project with ora2pg --project_base and run SHOW_REPORT --estimate_cost for assessment.
- Export TABLE, SEQUENCE, VIEW, TRIGGER, FUNCTION, PROCEDURE and PACKAGE types and load DDL.
- Export data with TYPE=COPY using parallel jobs; review DEFAULT_NUMERIC and DATA_TYPE overrides first.
- Run TEST and TEST_COUNT exports to compare object and row counts.
SQL Server Migration Assistant for Oracle partial effort M
SSMA replaces the schema half of Data Pump by reading metadata from a live Oracle connection, converting it and synchronizing it to SQL, and replaces the data half with its own migration engine that copies table data client-side or server-side through the extension pack. It cannot read .dmp files and does not offer Data Pump's network-mode or parallel worker tuning, so very large data sets are usually moved with Data Factory instead.
- Connect SSMA to the source Oracle instance and select the schemas to assess and convert.
- Synchronize the converted schema to the target, then run Migrate Data per table or schema.
- Install the extension pack on a SQL Server staging instance for server-side migration of large tables when the target is Managed Instance or SQL Server.
- Validate with SSMA Tester or row counts.
Azure SQL Managed Instance workaround effort M
Data Pump dump files are Oracle-proprietary and nothing on the SQL side can read them; Managed Instance restores only native SQL Server backups from Blob Storage. Data therefore moves with SSMA's data migration, Azure Data Factory's Oracle connector, or flat files landed in Blob Storage and loaded with BULK INSERT or OPENROWSET, with Data Pump useful only for producing external-table extracts or refreshing a staging Oracle copy.
- Use SSMA server-side or client-side data migration for initial loads after schema conversion.
- For large tables, extract with Data Factory using partitioned parallel copy into Parquet or CSV.
- Load with BULK INSERT from an external data source of type BLOB_STORAGE.
- Use Data Pump only to stage an Oracle copy for the migration run, never as an import format.
Notices of correction
- oracle Overview of Oracle Data Pump (Utilities 19c) · Database Utilities 19c · read 2026-09-26
- oracle Oracle Database Changes, Desupports, and Deprecations (Upgrade Guide 26ai) · Upgrade Guide 26ai, G43570-12, updated Jul 2026 · read 2026-09-26
- microsoft Overview: Oracle Database@Azure · page updated 2026-09-15 · read 2026-09-26
- microsoft Oracle Database@Azure: Exadata services · page updated 2026-04-15 · read 2026-09-26
- microsoft Azure Database Migration Service: supported migration scenarios · page updated 2026-02-20 · read 2026-09-26
- microsoft Azure Database Migration Service documentation · landing page updated 2026-08-18 · read 2026-09-26
- microsoft Migration service in Azure Database for PostgreSQL flexible server · page updated 2026-06-24 · read 2026-09-26
- community Ora2Pg documentation (configuration directives) · Ora2Pg 25.0 (changelog 2025-04-20); text read from doc/Ora2Pg.pod in the GitHub repository · read 2026-09-26
- microsoft SQL Server Migration Assistant for Oracle (OracleToSQL) · page updated 2026-04-24 · read 2026-09-26
- microsoft Oracle to Azure SQL Database: migration guide · page updated 2025-11-25 · read 2026-09-26
- microsoft T-SQL differences between SQL Server and Azure SQL Managed Instance · page updated 2026-09-09 · read 2026-09-26
- microsoft Oracle to Azure SQL Managed Instance: migration guide · page updated 2026-08-27 · read 2026-09-26
- microsoft Copy data to and from Oracle with Azure Data Factory · page updated 2026-08-05 · read 2026-09-26
- microsoft Bulk access to data in Azure Blob Storage (BULK INSERT and OPENROWSET) · page updated 2025-02-27 · read 2026-09-26