v5.6.4

5.6.4 (2026-04-16)

Fixes:

• Fixed upsert matching for relationship external IDs such as Account__r.Name when the source is csvfile and the target org returns the parent relationship as a nested Salesforce API object. Existing target child records are now matched and updated instead of being treated as new inserts.

See the related articles:

• ScriptObject Object - externalId
• Composite External Id Keys

v5.6.3

5.6.3 (2026-03-05)

Fixes:

• Fixed CSV null-token handling so bare N/A is preserved as a literal value while #N/A remains a null marker.
• Fixed lookup query generation for child objects with SELECT all so nested parent relationship segments from composite external IDs are not leaked as invalid bare relationship fields.

See the related articles:

• Values Mapping - check #N/A usage for null resets.
• Composite External Id Keys
• ScriptObject Object - externalId

v5.6.2

5.6.2 (2026-02-25)

Fixes:

• Fixed core:ExportFiles query parsing for polymorphic fields (for example ParentId$Account) so initialization keeps original polymorphic field names in composed queries.

See the related articles:

• ExportFiles Core Add-On Module
• Polymorphic Lookup Fields
• Advanced Examples - review FeedItem and polymorphic query scenarios.

v5.6.1

5.6.1 (2026-02-24)

Fixes:

• Fixed useSourceCSVFile processing for org -> org object-set runs: source CSV preparation now executes for CSV-based source tasks even when the source org is not csvfile.
• Fixed source CSV cleanup timing so prepared object-set source files are rebuilt deterministically before copy/repair processing.

Other changes:

• Updated Dependabot ignore rules for blocked advisories (ajv, minimatch, fast-xml-parser) to keep dependency update noise aligned with current project constraints.

See the related articles:

• ScriptObject Object - useSourceCSVFile
• Use Separate CSV Files

v5.6.0

5.6.0 (2026-02-22)

New features:

• Added --file CLI flag to run a migration from an explicit export.json path while keeping --path as the base folder for runtime resources.
• Added diagnostic lines that show the resolved export.json path and a dedicated note when this path differs from --path.
• Added anonymisation support for explicit export.json path values in diagnostic logs when --anonymise is enabled.

Improvements:

• Improved stack trace anonymisation to mask only the absolute machine-specific prefix while keeping plugin-root relative file paths visible.

Fixes:

• Updated non-zero exit guidance text to use consistent team wording in the final diagnostic instruction line.

See the related articles:

• Running - --file
• Run Command Flags - file
• Run Command Flags - path
• Run Command Flags - diagnostic
• Run Command Flags - anonymise
• Log File Management - What is masked and what is not

v5.5.0

5.5.0 (2026-02-22)

New features:

• Added skipped-update reason breakdown counters in warning output for update and upsert processing: sameData, noMatchingTarget, and other.
• Added detailed skipped-update summary data to make investigation of update/upsert behavior faster in large runs.

Fixes:

• Fixed missing diagnostic visibility for skipped-update split details by adding a dedicated diagnostic summary line in execution logs.

See the related articles:

• Log File Management in Plugin Migrations - Skipped Records Reason Breakdown
• Run Command Flags - diagnostic
• ScriptObject Object - skipRecordsComparison
• Target Records Filter review skipped updates when target rows are filtered out.

v5.4.0

5.4.0 (2026-02-21)

New features:

• Added object-level API mode controls: alwaysUseRestApi, alwaysUseBulkApi, and alwaysUseBulkApiToUpdateRecords.
• Added respectOrderByOnDeleteRecords to enforce ordered delete execution through REST one-by-one mode.
• Added SDK contract updates for add-on modules to expose new object-level API mode fields and related runtime metadata.

Improvements:

• Changed query and DML bulk-threshold behavior to threshold-inclusive evaluation (>=) for clearer switching at exact threshold values.

See the related articles:

• ScriptObject Object - check new object-level API mode properties and delete-order behavior.
• Script Object - scan bulk/query threshold settings and runtime switching behavior.
• Custom Add-On API Object Reference - go through SDK contract fields available to add-on modules.
• ScriptObject Object - alwaysUseRestApi
• ScriptObject Object - alwaysUseBulkApi
• ScriptObject Object - alwaysUseBulkApiToUpdateRecords
• ScriptObject Object - respectOrderByOnDeleteRecords

v5.3.1

5.3.1 (2026-02-20)

Fixes:

• Fixed composite externalId handling for mixed relationship and local fields (for example Lookup__r.Name;LocalField__c), so query expansion and externalId validation resolve correctly without false mandatory externalId warnings.

See the related articles:

• ScriptObject Object - externalId
• Export.json File Overview - look at where externalId is configured in migration scripts.
• ScriptObject Object - query
• Script Object - apiVersion

v5.3.0

5.3.0 (2026-02-20)

New features:

• Added automatic default apiVersion selection when apiVersion is not explicitly provided in export.json or via --apiversion.
• For org -> org migrations, SFDMU now uses the lower maximum API version supported by source and target orgs.
• For org -> csvfile and csvfile -> org migrations, SFDMU now uses the maximum API version supported by the connected org.
• Added support for the same auto-selection behavior when org credentials are provided manually with instanceUrl and accessToken.
• Kept explicit apiVersion in export.json and --apiversion as strict overrides over auto-selection.

See the related articles:

• Script Object - apiVersion
• Running - --apiversion
• ScriptObject Object - query

v5.2.0

5.2.0 (2026-02-20)

New features:

• Added --failonwarning to stop migration on the first warning and return dedicated exit code 8.
• Added mockFields.locale so mock data generation can use locale-specific values per field.
• Added Old Id in _target.csv reports, mapped from internal source id, for easier source-target traceability.

Fixes:

• Fixed warning escalation flow to always write an explicit diagnostic stop reason when --failonwarning aborts execution.
• Fixed diagnostic stack traces to mask absolute folder paths while preserving file names and line numbers.

Improvements:

• Improved diagnostic reporting for exclusion paths with detailed per-field reasons for operation-specific DML filtering.
• Improved Person Account diagnostics by logging exclusion reasons for invalid fields in business and person contexts.

See the related articles:

• Run Command Flags - failonwarning
• Running - --failonwarning
• MockField Object - locale
• Data Anonymization - look at patterns, examples, and usage guidance.
• The Target CSV Files - verify Old Id mapping output in target report files.
• Script Object - createTargetCSVFiles
• ScriptObject Object - query
• Script Object - apiVersion