SQLite identifier quoting
QuoteMica supports project-based AI conversations about database naming code. Use a compatible model endpoint to examine SQLite identifier boundaries, then review and test the proposed quoting function in your own environment.
def quote_identifier(name):
if not isinstance(name, str) or not name or chr(0) in name:
raise ValueError("identifier.invalid")
return '"' + name.replace('"', '""') + '"'- Connection
- Base URL, credentials and a reachable server.
- Project
- Check spaces and SQL-looking text in the name, rejecting empty names and NUL characters without treating a dotted name as multiple identifiers.
- Review before applying
- Compare actual behavior with the expected result. If a check fails, investigate it before applying more changes.
def quote_identifier(name):
if not isinstance(name, str) or not name or chr(0) in name:
raise ValueError("identifier.invalid")
return '"' + name.replace('"', '""') + '"'Separate a query's structure from the values it carries
Dynamic query code needs a clear distinction between a value to search for and the name of a database object. QuoteMica provides a Windows AI coding workspace for reviewing that distinction in your project's files. Start with the query-producing function and explain which parts are allowed to vary. Ask the assistant to trace where each name and value originates before proposing edits.
The SQLite identifier-quoting example is one case you could examine. Review how quoted names reach the query builder and whether the application should accept those names in the first place. Correct representation and authorization are separate decisions. A useful proposal explains both the helper's limited responsibility and the caller's constraints, then identifies concrete queries you can execute to check the intended behavior.
Choose an uncensored AI model if your supported OpenAI-compatible or Anthropic endpoint offers one, adding credentials as required. Its provider policies, tool support and context limits still govern the connection. Read the proposed query changes and run your project's database tests against representative schema names. A local model uses a separately running server and does not establish that database tools or other connected services operate locally.
Review points for dynamic query code
Does quoting a table name mean it should be allowed?
No. Quoting represents an identifier in query syntax; the application still needs to decide which objects may be addressed. Ask for inspection of that decision at the caller, especially when a name comes from user-controlled input.
Can I handle column names like ordinary bound values?
Treat query structure and data values as separate review subjects. Inspect what your database interface supports and how the existing query is built. The identifier example is not evidence that binding a value can select an arbitrary column name.
What should a query-builder regression check demonstrate?
It should exercise the actual query with an allowed name that needs quoting and confirm the intended result. Include rejected names where the application restricts them. Checking the produced string alone does not verify execution or authorization behavior.
Will QuoteMica discover every dynamic query automatically?
Do not assume complete coverage. Ask which files and call sites informed the proposal, and identify any known query builders yourself. Model context and available tools limit inspection; review unexamined paths before treating a broader change as complete.
QuoteMica · Updated 2026-10-25