Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Importing existing CAD folder trees

Question: how do PLM/PDM systems bring an existing folder of CAD data – often hundreds of files, with project files – into the vault, and what should TeamCraft’s import do? Investigated on 2026-11-04.

How other systems do it

SystemToolWhat it does
Autodesk VaultAutoloader, or Check In Project in Inventorscans the folder with the Inventor project file (.ipj), reports missing references, uploads in batches (up to 20 000–30 000 Inventor files), keeps the folder structure, works on a copy – the source files stay untouched; DWF creation can be switched off for speed
SolidWorks PDMcopy/paste of folders, scan tools, VAR import programsmigration guides: scan first, map folders and properties, fix broken references and duplicates before the import, decide whether files come in at their existing revision
Teamcenter with Solid EdgeSEEC Data Preparation: Analyze, then Add to Teamcenter (with dry run)expects a flat source folder without duplicate file names; Analyze reports duplicate names, duplicate document numbers and broken links; item IDs are generated when the document number is empty; Design Manager can flatten folders and rename duplicates

The common pattern:

  1. Prepare – one project per product (Inventor: gather the files and the .ipj into one project, e.g. with Pack and Go), fix broken references, make file names unique.
  2. Analyse first (dry run) – a report of what will be imported, duplicate names, broken references; nothing is changed yet.
  3. Import – in batches, from a copy (the source stays untouched), with generated or property-based numbers; viewables optional.
  4. Afterwards – check the structure, release the data through the normal workflow.

Decision for TeamCraft

TeamCraft matches CAD references by file name, like Teamcenter – so file names must be unique within an import. The desktop app gets an Import folder… wizard:

  • Scan (dry run): walks the folder, skips lock/backup files and OldVersions folders, sorts the files into models (parts, assemblies), drawings, exports (STEP, STL, PDF, …), project files (.ipj, …) and other files, and reports duplicate file names. Nothing is uploaded yet.
  • Plan: one item per model file (assembly or part); a drawing and exports with the same base name go into the same item (the item’s datasets); project files and other documents can go into one document item named after the folder. Numbers are generated, or taken from the file names when they are valid item numbers.
  • Import: item by item with progress and stop, from the original files (read only); every new item is checked in at the end, so nothing stays locked; the server scans the assemblies for the files they reference.
  • Structure: optionally BOM lines from the references found in the assemblies, with the instance count the server read from the assembly (instances of GET …/references: SOLIDWORKS 2015+ component tree, NX structure); where the format does not tell (Inventor, Solid Edge, …) quantity 1, to be checked – exact quantities come from the BOM export of the CAD program (Import BOM…).
  • Release afterwards with one change order, like every other release in TeamCraft (no import “as released” past the lifecycle).

An in-app tutorial (desktop Help → Import a CAD folder) explains the preparation (Pack and Go, project file, unique names), the wizard and the check afterwards.

Sources

Status

  • Implemented in the desktop app: Tools → Import CAD Folder… (desktop/src/folderimport.rs for the scan, the plan and the import in desktop/ui/app.js) and the tutorial Help → How to Import a CAD Folder. Imported items are remembered per folder (browser storage of the app), so an interrupted import continues where it stopped.
  • Tested: the folder scan (unit test) and the planning with a sample tree (items, duplicates, Creo iterations, project files). Not yet tried with a real Inventor or Solid Edge project of hundreds of files – the BOM lines depend on the heuristic reference scan of the server.