One clarification from my above post. With line data, you would use seeds to create polygons in the parcel fabric. With polygon data, you can use the Append and Build Parcels tools, which will create parcel fabric polygons, lines and points.
Records exist for a few reasons: one, to link individual parcels to their corresponding recorded legal description; two, to link parcels with a parcel type in the fabric, and; three, for conformity with parcel fabric topology rules. Records can be created (and persisted) without a geometry--in other words, we can add a new record to the fabric called "New_Subdivision" without associating it to vector data. To that end, "the record footprint or polygon matches the cumulative geometry of all the parcels associated with it."
Without knowing more about your particular situation, here's a hypothetical example: Let's say we have a subdivision that's being installed in multiple phases, yet the entire development has the same record on a legal description with the county recorder. The first phase is complete and has been added to our parcel fabric, and now we need to update the fabric with polygon data for the second phase of development. Because the legal record on file with the county recorder is the same for both phases, we want to add the parcels of the second phase to an existing record. In this example, we would:
- append second phase polygons to the parcel type,
- build parcels;
- update COGO (if needed)
- assign parcels to the record,
- validate and fix topology errors (if needed)
Once we have assigned parcels to a record--in this case the same record as the first phase of development--the record feature will change in size to include the second phase of the development.
Are you migrating your fabric from an ArcMap fabric? If so, there's a slightly different workflow: Upgrade an ArcMap parcel fabric—ArcGIS Pro | Documentation
Here are a couple more resources about the parcel fabric data model and parcel records: