getredash/redash · error · ValueError

Queries of type {} can not be processed by redash.

Error message

Queries of type {} can not be processed by redash.

What it means

CorporateMemoryRunner only accepts SPARQL SELECT queries (or queries whose type can't be determined, which are attempted anyway). SparqlQuery.get_query_type returns ASK/CONSTRUCT/DESCRIBE for other forms, and those are rejected before execution.

Source

Thrown at redash/query_runner/corporate_memory.py:158

    @classmethod
    def enabled(cls):
        return enabled

    @classmethod
    def type(cls):
        return "corporate_memory"

    def run_query(self, query, user):
        """send a sparql query to corporate memory"""
        query_text = query
        logger.info("about to execute query (user='{}'): {}".format(user, query_text))
        query = SparqlQuery(query_text)
        query_type = query.get_query_type()
        # type of None means, there is an error in the query
        # so execution is at least tried on endpoint
        if query_type not in ["SELECT", None]:
            raise ValueError("Queries of type {} can not be processed by redash.".format(query_type))

        self._setup_environment()
        try:
            data = self._transform_sparql_results(query.get_results())
        except Exception as error:
            logger.info("Error: {}".format(error))
            try:
                # try to load Problem Details for HTTP API JSON
                details = json.loads(error.response.text)
                error = ""
                if "title" in details:
                    error += details["title"] + ": "
                if "detail" in details:
                    error += details["detail"]
                    return None, error
            except Exception:
                pass

View on GitHub (pinned to ca79fe988d)

Solutions

1. Rewrite the query as a SELECT with the projections you want
2. Convert CONSTRUCT patterns to SELECT ?s ?p ?o form
3. Express boolean checks as SELECT (COUNT(*) AS ?n) or SELECT (bound(?x) AS ?result)

Example fix

# before
ASK { <http://ex/x> ?p ?o }

# after
SELECT (COUNT(*) AS ?triples) WHERE { <http://ex/x> ?p ?o }

if not query_text.lstrip().upper().startswith('SELECT'):
    raise ValueError('rewrite as a SELECT projection before sending to Redash')

def is_select_sparql(q: str) -> bool:
    return q.lstrip().upper().startswith('SELECT')

try:
    data, err = runner.run_query(q, u)
except ValueError as e:
    if 'can not be processed' in str(e):
        q = to_select_projection(q); retry

When it happens

Trigger: Sending a SPARQL query of type ASK, CONSTRUCT, DESCRIBE, or any other non-SELECT form to a Corporate Memory data source.

Common situations: Reusing SPARQL from tools that default to CONSTRUCT for graph extraction, or expecting Redash to tabulate non-SELECT results.

Related errors

• Queries of type {} can not be processed by redash.
• Invalid JWT token
• Error during query execution. Reason: {error}
• Failed running query [%s].
• Username and Password required

AI-assisted analysis of getredash/redash@ca79fe988d (2026-08-28). Data as JSON: /api/errors/ebb6c9f73886b608. Report an issue: GitHub.