I absolutely agree that truncate and truncation are overloaded terms on their own, even when limiting their usage strictly to databases. However, should these tags really just go? It may still be useful to have log truncation and table truncation, as concepts, covered by tags, and if we agree on that, why burninate the tags when we could make them synonyms of more accurately named tags?
The truncate tag has already got some usage guidance:
A DDL operation that quickly removes all rows.
Based on that, I suggest we keep it as a synonym of a new tag called sql-truncate – or, perhaps, it could be truncate-table.
And for log truncation we could create a tag called log-truncation and make truncation its synonym.
As for numeric truncation and string truncation, the former is too generic an operation to need a dedicated tag in the context of working with databases, while the latter can be viewed as part of the variety of string operations that are sufficiently covered by the already existing string-manipulation tag. Still, we could add a synonym, string-truncation, for it if we agree that this type of operation comes up often enough in questions that we should enable the system to guide a question asker or a tag editor to the correct tag.