MD5 / SHA-1 / SHA-256 / SHA-512 of a string or a file - verify the source before relying on it.
The tools that come up when the actual job is interpreting data rather than collecting it: CSV viewer with column typing for the export you just got from the database, JSON formatter for the API response, regex tester against the source string, base64 for the binary blob someone embedded in a JSON field, hash for source verification, and the AI summarizer for digesting long-form sources.
SHA-256. If they publish MD5 or SHA-1 only, you can still catch an accidental corruption, but you cannot distinguish that from a deliberately altered file - which is the point of the check.
The bytes are not identical: line-ending conversion, a re-save or an added header all change the hash. Transfer in binary mode, then hash both sides before suspecting the data itself.
No. Hashing is one-way but deterministic, so the same input always produces the same output - it is pseudonymisation, not anonymisation. If you need to aggregate rather than hide, decide that before hashing.