[FEATURE] Preserve identifier type information in anonymizer

Is your feature request related to a problem?
For data confidentiality purposes, our anonymizer replaces all field names with identifier:

source=table | where identifier > identifier | stats count() by identifier

This gets the job done, but makes debugging queries from log entries difficult. For reproducing certain types of queries from bugs, it would be nice to at least have type information.

What solution would you like?
Preserve type information in the identifier output, turning the above query into something like:

source=table | where field_integer > field_integer | stats count() by field_string

This is considerably more actionable for reproducing bugs from logged queries.

What alternatives have you considered?
Don't

Do you have any additional context?
N/A

Start by locating the anonymizer implementation and tracing how field names become identifier in the SQL/PPL query examples. Determine how identifier types can be retained without exposing field names, then verify that anonymized queries preserve examples such as integer comparisons and string grouping while remaining actionable for logged-query debugging.