mem0ai/mem0 · error · ValueError

Invalid filter key: {key!r}

Error message

Invalid filter key: {key!r}

What it means

BaiduDB._create_filter raises ValueError when a metadata filter key fails the _SAFE_FILTER_KEY regex ^[a-zA-Z_][a-zA-Z0-9_]*$. Because keys are interpolated directly into a SQL-like expression metadata["key"] = value, non-identifier keys (spaces, dots, hyphens, leading digits, empty strings) are rejected to prevent injection into the Baidu query language.

Source

Thrown at mem0/vector_stores/baidu.py:418

            )
        except Exception as e:
            logger.warning(f"Error resetting table: {e}")
            raise

    def _create_filter(self, filters: dict) -> str:
        """
        Create filter expression for queries.

        Args:
            filters (dict): Filter conditions.

        Returns:
            str: Filter expression.
        """
        conditions = []
        for key, value in filters.items():
            if not self._SAFE_FILTER_KEY.match(key):
                raise ValueError(f"Invalid filter key: {key!r}")
            if isinstance(value, str):
                escaped = value.replace("\\", "\\\\").replace('"', '\\"')
                conditions.append(f'metadata["{key}"] = "{escaped}"')
            elif isinstance(value, (int, float, bool)):
                conditions.append(f'metadata["{key}"] = {value}')
            else:
                raise ValueError(
                    f"Filter value for {key!r} must be str, int, float, or bool, "
                    f"got {type(value).__name__}"
                )
        return " AND ".join(conditions)

View on GitHub (pinned to 001c235229)

Solutions

1. Rename the metadata key to a plain identifier (letters, digits, underscore, not starting with a digit) both in stored payloads and in filters.
2. Sanitize keys before calling search: re.sub(r'[^A-Za-z0-9_]', '_', key).
3. Drop or reject non-conforming keys early at the API boundary instead of letting the vector store throw.

Example fix

# before
results = db.search(query="...", vectors=[...], filters={"user-id": "alice"})  # ValueError

# after
results = db.search(query="...", vectors=[...], filters={"user_id": "alice"})

import re
_SAFE_KEY = re.compile(r"^[a-zA-Z_][a-zA-Z0-9_]*$")

def sanitize_filters(filters: dict) -> dict:
    out = {}
    for k, v in (filters or {}).items():
        if not _SAFE_KEY.match(k or ""):
            k2 = re.sub(r"[^A-Za-z0-9_]", "_", str(k))
            if not re.match(r"^[A-Za-z_]", k2):
                k2 = "f_" + k2
            k = k2
        out[k] = v
    return out

db.search(query, vectors, filters=sanitize_filters(filters))

def has_safe_filter_keys(filters: dict) -> bool:
    import re
    return all(isinstance(k, str) and re.match(r"^[a-zA-Z_][a-zA-Z0-9_]*$", k) for k in (filters or {}))

When it happens

Trigger: Calling search/get on BaiduDB with filters={'user-id': 'x'}, {'meta.key': 'v'}, {'2key': 1}, or any key containing characters outside [A-Za-z0-9_] or starting with a digit. Only runs when filters is a non-empty dict.

Common situations: Passing structured metadata keys that came from JSON payloads with dots/hyphens; copying filter syntax from another provider (e.g. Qdrant nested 'metadata.field' paths); machine-generated keys from user data.

Related errors

• Invalid filter key: ${key}
• Filter value for ${key} must be str, int, float, or bool, go
• Invalid filter key: ${JSON.stringify(key)}
• Invalid filter key: ${JSON.stringify(key)}
• Filter list for '${key}' contains an object, which may conta

AI-assisted analysis of mem0ai/mem0@001c235229 (2026-08-15). Data as JSON: /api/errors/8596f53551843744. Report an issue: GitHub.