Literal SQL LIKE patterns
LikeSail is a Windows AI coding client for working through query behavior with your own project context. Review literal-pattern escaping with a compatible model and keep the bound value separate from the SQL statement when applying its suggestions.
def like_literal(text):
if type(text) is not str or "\x00" in text:
raise ValueError("like.text")
return text.replace("!", "!!").replace("%", "!%").replace("_", "!_")
SQL = "SELECT value FROM items WHERE value LIKE ? ESCAPE '!'"- Connection
- Base URL, credentials and a reachable server.
- Project
- Test percent, underscore, exclamation marks and empty text with a bound parameter, rejecting NUL without adding wildcards or claiming case-sensitive matching.
- Review before applying
- Compare actual behavior with the expected result. If a check fails, investigate it before applying more changes.
def like_literal(text):
if type(text) is not str or "\x00" in text:
raise ValueError("like.text")
return text.replace("!", "!!").replace("%", "!%").replace("_", "!_")
SQL = "SELECT value FROM items WHERE value LIKE ? ESCAPE '!'"Make the search box mean what its users type
A person entering punctuation into a search field may expect those characters to be found literally. The database query may interpret them differently. LikeSail is a Windows AI coding workspace where you can investigate that mismatch in your application, from the input field to the query call. Describe the intended search behavior and ask for inspection of the code that constructs and executes it.
The SQL LIKE example offers a focused development task around percent signs, underscores and an escape character. You could ask the assistant to review how literal text becomes a bound query value and how the query declares its escape convention. Check the caller as well as the helper, then test actual database results. Keep literal matching requirements distinct from broader questions about query safety and permitted search syntax.
For abliterated AI model options, choose an available model through a supported OpenAI-compatible or Anthropic endpoint and supply required credentials. The endpoint's model, server and provider determine context, tools and restrictions. Review any proposed edits before running your project's database checks. If you use a local model server, other connected tools and the database connection can still have their own external destinations.
Defining the search contract
- Should percent and underscore always be treated literally?
That depends on the search interface you intend to offer. State whether people can enter wildcard expressions or only ordinary text. Inspect the query construction against that decision, and test punctuation rather than assuming the displayed input keeps its meaning.
- Is escaping LIKE characters a substitute for bound parameters?
No. Literal-pattern handling and query value binding solve different concerns. Review how the helper's result reaches the database call and preserve the intended parameterization. Do not infer broad query safety from a small escaping example alone.
- Which database examples should accompany the change?
Include rows containing the wildcard characters, the escape character and ordinary text, with expected matches for each search. Execute them against the project's database setup so the query and helper are tested together under the same convention.
- Does LikeSail change the database search settings for me?
No automatic configuration change is implied. Ask for inspection of the existing query and connection behavior, then review any proposed code or configuration edits. Your project's actual database remains the place to verify the resulting search semantics.
LikeSail · Updated 2026-10-25