Skip to content

Import Existing Data

Import is most useful for moving an existing register into a governed model. Preview the generated SysML before writing files.

1. Inspect available kinds

memo-architect ontology show
memo-architect import template elements --output elements-template.csv
memo-architect import template relationships --output relationships-template.csv

2. Map source columns

Prepare stable IDs, names, kinds, descriptions, and relevant attributes:

id,name,kind,statement,priority
NEED-001,Safe therapy,Need,Patient needs safe delivery,high
REQ-001,Flow accuracy,Requirement,The device shall deliver within the specified tolerance,high

Kind names must exist in the active profile. Preserve source identifiers when they are stable and unique.

3. Preview

memo-architect import csv requirements.csv --dry-run

Review:

  • element kinds;
  • sanitized SysML usage names;
  • escaped text;
  • package name;
  • attributes that could not be mapped.

4. Generate source

memo-architect import csv requirements.csv \
  --output model/requirements/imported.sysml \
  --package imported_requirements

5. Import relationships

sourceId,targetId,type
NEED-001,REQ-001,DerivesFrom

memo-architect import csv-rel traceability.csv \
  --output model/traceability/imported.sysml \
  --package imported_traceability

Validate immediately. Importing rows creates syntax; it does not prove that the selected kinds and relationships are correct.

6. Keep a model in step with another system

import csv writes elements once. When a requirements tool, a PLM or a test system stays the owner of the records, import them as a source instead, so the next export can be compared with what the model holds:

memo-architect import source requirements.csv --system reqtool --kind Requirement
memo-architect import source export.reqif --system reqtool --format reqif --kind Requirement
memo-architect import source trace.csv --system reqtool --format trace-matrix

A source's records go in one file, model/imports/<system>.sysml, that the import owns and regenerates. Do not edit it; put your own elements and links elsewhere.

  • Source id. The CSV needs a sourceId (or id) column and may have revision; ReqIF uses IDENTIFIER and LAST-CHANGE. Each record becomes an element with a name MEMO derives (REQ-CS-01 is reqCs01) and an @ExternalSource naming the system, the source id and the revision. A record with no revision gets one from its content. A name is kept for as long as the source id exists, so links to the element keep working.
  • Re-import. Run the same command on the new export. It lists what is new, what changed and what was deleted, each with the links that touch it, and writes the file unless you pass --dry-run.
  • Deleted. A record that left the export is kept and marked sourceWithdrawn; it is never removed, so the links you wrote to it still compile. A record that comes back is no longer withdrawn.
  • Suspect links. A link you author to an imported element can record the revision it was made against, in its body: satisfy req1 by pump { @TraceBaseline { :>> sourceRevision = "B"; } }. The link is suspect once the element's revision is no longer that one, or the element is withdrawn. Nothing stores that status; it is derived each time.
  • Traces from the source. A trace matrix (from,to,type) or a ReqIF relation is written inside the source's file, without a baseline, because the source owns it. An export of elements alone leaves those traces in place; one that carries traces replaces them.

A column the kind does not declare is not written, and the command says which.