Account mapping rule configuration is not applied to file based runs. Offline runs are processed in a shared service context rather than in your account context, so mapping rules configured against your account do not take effect. The same rooms can therefore group differently in a file based run than they do through the API. If you depend on your own mapping rules, use the real time or asynchronous path.
No confidence scores. The output file carries the mapping result only. There is no matchScore and no cfs, so you cannot rank or triage the results by confidence the way you can on the API.
Rows beyond the batch limit are skipped silently. A maximum of 100,000 rooms are processed per batch. Rows beyond that are skipped, and the run still finishes with status Completed. Always compare the row count of your output file against your input file.
Column headers are matched exactly, and letter case matters. A misspelled or differently cased header is not reported as an error. The column is read as empty instead, and the file is processed as though you had sent no value at all.
A run is all or nothing. If one hotel in the file cannot be processed, the whole run fails and no partial output is uploaded. You resubmit the entire file.
There is no idempotency key. Re-uploading the same file starts a new and independent run. Nothing detects or merges duplicate submissions.
There is no per row error report. The only failure detail is a run level status and reason, so a problem with a single row is not attributed to that row.
Supply VervotechHotelId wherever you have it. When a hotel cannot be resolved to a Vervotech ID, its rooms are grouped only among themselves and carry a placeholder identifier in the output, which cannot be joined to your Vervotech content.
File based runs are processed in English. The culture option is not available on this path.
Choose real time for live pricing and availability flows, for anything that needs a confidence score, and whenever your own mapping rules must apply.
Choose asynchronous when a single hotel carries more rooms than the real time path accepts and you can tolerate polling for the result.
Choose file based for one off or periodic bulk analysis across many hotels, for catalogue onboarding, and for evaluating coverage before you build an integration.
