fix(mcp-kb): SQL HAVING clause without GROUP BY breaks semantic search #73

New Issue

cardosofelipe · 2026-01-03T23:57:01Z

cardosofelipe commented

2026-01-03 23:57:01 +00:00

Bug Description

In mcp-servers/knowledge-base/database.py, the _format_where_clause method uses HAVING without a GROUP BY clause, which breaks PostgreSQL queries when semantic search is combined with filters.

Location

mcp-servers/knowledge-base/database.py:393-395

# BUG: HAVING without GROUP BY
if min_score:
    having_clauses.append(f"1 - (embedding <=> $1) >= {min_score}")

Expected Behavior

The query should properly filter by minimum similarity score without breaking PostgreSQL syntax.

Actual Behavior

Queries with min_score parameter will fail with PostgreSQL syntax errors.

Fix Required

Either:

Add proper GROUP BY clause when using HAVING
Use a subquery/CTE approach
Move the condition to WHERE with proper vector distance calculation

Found during PR #72 deep review
Issue #57 (Knowledge Base MCP Server)

## Bug Description In `mcp-servers/knowledge-base/database.py`, the `_format_where_clause` method uses `HAVING` without a `GROUP BY` clause, which breaks PostgreSQL queries when semantic search is combined with filters. ### Location `mcp-servers/knowledge-base/database.py:393-395` ```python # BUG: HAVING without GROUP BY if min_score: having_clauses.append(f"1 - (embedding <=> $1) >= {min_score}") ``` ### Expected Behavior The query should properly filter by minimum similarity score without breaking PostgreSQL syntax. ### Actual Behavior Queries with `min_score` parameter will fail with PostgreSQL syntax errors. ### Fix Required Either: 1. Add proper `GROUP BY` clause when using `HAVING` 2. Use a subquery/CTE approach 3. Move the condition to `WHERE` with proper vector distance calculation ### Related - Found during PR #72 deep review - Issue #57 (Knowledge Base MCP Server)

cardosofelipe closed this issue

2026-01-04 00:03:42 +00:00

cardosofelipe referenced this issue from a commit

2026-01-04 00:08:59 +00:00

fix(mcp-kb): address critical issues from deep review

~~cardosofelipe referenced this issue 2026-01-04 00:50:16 +00:00~~

feat(mcp): Context Management Engine #61

Sign in to join this conversation.