Microsoft Project XML Import and Export

Importing from Microsoft Project

To prepare an XML file for import, open your project in Microsoft Project and choose File > Save As, then select XML Format (.xml) as the file type. The resulting file follows the MSPDI schema and can be imported into Maverick immediately.

Maverick's import layer accepts the XML in three ways, giving you flexibility depending on where the file comes from:

• ImportFromFile — reads the XML directly from a file path on disk. Use this from the Windows desktop application when you have the .xml file saved locally.
• ImportFromBytes — accepts the XML as a UTF-8 byte array. Use this when streaming an uploaded file through the web interface without saving it to disk first.
• ImportFromXml — accepts the XML as a string. Use this when the XML has already been read or generated in memory. The other two methods delegate here internally.

When you supply an existing Maverick project GUID, the import updates that project's header fields and tasks. When you supply an empty string for the project ID, Maverick creates a brand-new project from the MSPDI <Name> element and returns its GUID.

Re-import safety

Importing the same XML file a second time does not create duplicates. The importer uses a three-step lookup for each task it encounters:

1. Match by MSPDI UID — the integer <UID> stored in the database as MSTaskID from the first import. This is the fastest and most reliable match.
2. Match by task name within the same project — catches tasks that were created in Maverick manually before the first import, which have no stored MSPDI UID yet.
3. Create a new record if neither lookup finds a match.

All three paths converge on the same field-update logic, so a first-time import and a re-import of the same file produce identical results. After the import completes, the critical path cache is invalidated so that CPM recalculates on the next access with the updated schedule data.

Exporting from Maverick

Maverick generates a complete MSPDI XML document that Microsoft Project opens without modification. The output is structured in five sections, in the order the MSPDI schema requires:

1. Project Header

The project name, start date, finish date, and notes are written first. The header also includes the work-time assumptions that control how Microsoft Project interprets all duration and lag values in the file: 480 minutes per day (8-hour work day), 2,400 minutes per week (5-day week), and 20 days per month.

2. Calendars

MSPDI requires at least one calendar element — Microsoft Project will reject a file that omits this section. Maverick emits a single Standard calendar. If you need custom working-time exceptions (company holidays, part-time schedules), those can be added after import inside Microsoft Project.

3. Tasks

Every task is emitted in display order with its outline level, WBS number, start, finish, duration, and percent complete. WBS numbers are computed automatically from the outline level sequence — you do not need to supply them. Milestone tasks omit the <Duration> and <Work> elements, which causes Microsoft Project to render them as diamond markers instead of bars. Each dependent task includes its <PredecessorLink> children with the link type code and lag value. A mandatory UID-0 project summary task is emitted first, as required by the MSPDI schema.

4. Resources

Each named worker registered via AddResource becomes a <Resource> element with a unique integer UID and a display name. A mandatory UID-0 null-resource placeholder is always emitted first. Resources are typed as work resources (<Type>1</Type>), which is the correct classification for named employees in Microsoft Project.

5. Assignments

Each task-to-resource assignment carries the task UID, resource UID, allocation units (1.0 = 100%, 0.5 = 50%), and planned work hours. If a resource has actual timesheet hours logged against a task, those are emitted as <TimephasedData Type="2"> children — one element per day — representing timephased actual work. Microsoft Project displays these in the Resource Usage view's timephased grid. If a worker logged actual hours against a task but was not on the planned assignment list, Maverick creates an implicit assignment with units set to 0 (tracking-only) so the timephased data has a valid parent element.

Resources, Assignments, and Actual Hours

On export, every named worker in the project becomes a <Resource> element with a unique integer UID. Resource names are deduplicated case-insensitively — two assignments to "Alice Smith" and "alice smith" produce a single resource record. The UID assignment follows insertion order, so UIDs remain stable across repeated exports of the same project.

Each task-to-resource pairing becomes an <Assignment> element. The allocation fraction is expressed as a decimal: 1.0 for a full-time assignment, 0.5 for half-time. Planned work hours for the assignment are derived as task.DurationHours × units — the same relationship Microsoft Project uses when you assign a work resource to a task.

Actual timesheet hours are emitted as timephased data inside each assignment. Each day that a resource logged time against a task produces one <TimephasedData Type="2"> element, bracketing the standard 8:00–17:00 work day for that calendar date. The total actual work across all days is also written to the <ActualWork> field on the assignment. Microsoft Project's Resource Usage and Task Usage views use this data to show how actual work is distributed over time.

If you need to export only a subset of tasks, you can call AddTask selectively — any tasks not added to the export instance simply will not appear in the file. The same applies to resources and assignments: only items you explicitly register are included. This allows you to generate a filtered snapshot — for example, all tasks assigned to a specific team — without modifying the underlying project data.