👤 User Story
As a node operator, I want export-dd to support namespace filtering and correct pagination so that I can export and inspect only the LDD entries for a specific namespace, and reliably export all entries when the total count exceeds the batch size.
💪 Motivation
export-dd currently:
- Accepts a
-ns <namespace> flag but silently ignores it — all documents are exported regardless
- Uses a placeholder
searchAfter = "" for pagination (marked FIXME in DataExporter.java:79) instead of extracting the last sort value from each response batch — meaning export silently stops at the first batch (~100 docs) when more exist
This makes export-dd unreliable as a diagnostic tool for inspecting what LDD field definitions are present in a registry node's -dd index.
📋 Requirements
export-dd -ns <namespace> should filter results to only documents belonging to the given namespace (i.e. where attr_ns matches)export-dd without -ns should export all documents, correctly paginating through all batches using search_afterexport-dd with -ns should also paginate correctly if a namespace has more than BATCH_SIZE entries
🩺 Additional context
DataExporter.export() in registry-mgr has a FIXME comment on the searchAfter assignment: searchAfter = ""; // FIXME: needs to be the last ID from the batch()Response.Search interface has no method to retrieve the last sort value from a response — one needs to be added- The AWS
SearchRespWrap and ES SearchRespImpl both need to implement the new method
- The
-ns filter should be passed through DDDataExporter.createRequest() → SearchImpl.all() using the existing all(filterField, filterValue, sortField, size, searchAfter) overload, filtering on attr_ns
🦄 Related requirements
For Internal Dev Team To Complete
⚙️ Engineering Details
🎉 Integration & Test
🤖 Generated with Claude Code
👤 User Story
As a node operator, I want
export-ddto support namespace filtering and correct pagination so that I can export and inspect only the LDD entries for a specific namespace, and reliably export all entries when the total count exceeds the batch size.💪 Motivation
export-ddcurrently:-ns <namespace>flag but silently ignores it — all documents are exported regardlesssearchAfter = ""for pagination (markedFIXMEinDataExporter.java:79) instead of extracting the last sort value from each response batch — meaning export silently stops at the first batch (~100 docs) when more existThis makes
export-ddunreliable as a diagnostic tool for inspecting what LDD field definitions are present in a registry node's-ddindex.📋 Requirements
export-dd -ns <namespace>should filter results to only documents belonging to the given namespace (i.e. whereattr_nsmatches)export-ddwithout-nsshould export all documents, correctly paginating through all batches usingsearch_afterexport-ddwith-nsshould also paginate correctly if a namespace has more thanBATCH_SIZEentries🩺 Additional context
DataExporter.export()inregistry-mgrhas aFIXMEcomment on thesearchAfterassignment:searchAfter = ""; // FIXME: needs to be the last ID from the batch()Response.Searchinterface has no method to retrieve the last sort value from a response — one needs to be addedSearchRespWrapand ESSearchRespImplboth need to implement the new method-nsfilter should be passed throughDDDataExporter.createRequest()→SearchImpl.all()using the existingall(filterField, filterValue, sortField, size, searchAfter)overload, filtering onattr_ns🦄 Related requirements
For Internal Dev Team To Complete
⚙️ Engineering Details
🎉 Integration & Test