Preserve the complete file set
- Copy data, keys/indexes, memo or blob companions, application binaries and configuration together.
- Identify file variants using signatures and application/compiler context rather than extension alone.
- Locate Clarion dictionaries, source, reports or screen labels that describe fields and relationships.
- Determine code page, date representation, numeric scale and deleted-record semantics.
- Use read-only or disposable copies for inspection and rebuilding indexes.
Missing keys may sometimes be rebuilt from intact records; missing memo companions can mean the variable-length content itself is unavailable. These are different recovery conditions and should not be reported as equivalent.
Reconstruct and validate
| Dictionary available | Verify it matches the deployed data version and actual record layout. |
|---|---|
| Dictionary missing | Infer schema from repeated records, code, UI, reports and known values. |
| Encoding uncertain | Test representative names and symbols against likely code pages. |
| Damaged file | Expect partial limits where headers, records or companion data were overwritten. |
Stage extracted data in explicit types, retain original identifiers and reconcile counts, totals and linked records before migration. Preserve uncertain fields rather than silently coercing them.
Use DBF recovery where the files are dBase-family, Pervasive SQL recovery for Btrieve/PSQL estates, format identification and the database pillar.