prestodb/presto · error · SemanticException

COLUMN_NAME_NOT_SPECIFIED

COLUMN_NAME_NOT_SPECIFIED

Error message

Column name not specified at position %s

What it means

During analysis, the relation descriptor's visible fields are checked to ensure every output column has a name. If any field's name is absent (Optional.empty()), analysis fails with COLUMN_NAME_NOT_SPECIFIED, reporting the 1-based ordinal position. Presto requires named columns for the result relation (e.g. for CREATE TABLE AS, views, named queries).

Source

Thrown at presto-main-base/src/main/java/com/facebook/presto/sql/analyzer/StatementAnalyzer.java:1614

            for (Property property : properties) {
                if (!propertyNames.add(property.getName().getValue())) {
                    throw new SemanticException(DUPLICATE_PROPERTY, property, "Duplicate property: %s", property.getName().getValue());
                }
            }
            for (Property property : properties) {
                process(property, scope);
            }
        }

        private void validateColumns(Statement node, RelationType descriptor)
        {
            // verify that all column names are specified and unique
            // TODO: collect errors and return them all at once
            Set<String> names = new HashSet<>();
            for (Field field : descriptor.getVisibleFields()) {
                Optional<String> fieldName = field.getName();
                if (!fieldName.isPresent()) {
                    throw new SemanticException(COLUMN_NAME_NOT_SPECIFIED, node, "Column name not specified at position %s", descriptor.indexOf(field) + 1);
                }
                if (!names.add(fieldName.get())) {
                    throw new SemanticException(DUPLICATE_COLUMN_NAME, node, "Column name '%s' specified more than once", fieldName.get());
                }
                if (field.getType().equals(UNKNOWN)) {
                    throw new SemanticException(COLUMN_TYPE_UNKNOWN, node, "Column type is unknown: %s", fieldName.get());
                }
            }
        }

        private void validateColumnAliases(List<Identifier> columnAliases, int sourceColumnSize)
        {
            if (columnAliases.size() != sourceColumnSize) {
                throw new SemanticException(
                        MISMATCHED_COLUMN_ALIASES,
                        columnAliases.get(0),
                        "Column alias list has %s entries but subquery has %s columns",
                        columnAliases.size(),

View on GitHub (pinned to 55bb57d202)

Solutions

1. Add an explicit alias to the unnamed output column: SELECT 1 + 1 AS result
2. If the SELECT is generated, ensure every non-trivial expression in the projection list has an AS clause
3. Re-run the statement; the error's position field tells which column (1-based) lacks a name

Example fix

// before
CREATE TABLE metrics AS SELECT now() - interval '1' DAY, count(*) FROM events;

// after
CREATE TABLE metrics AS SELECT now() - interval '1' DAY AS window_start, count(*) AS event_count FROM events;

// Before submitting DDL, ensure every projection item has an alias
List<String> projections = parseProjectionItems(sql);
for (int i = 0; i < projections.size(); i++) {
    if (!projections.get(i).matches(".*\\s+AS\\s+\\w+.*")) {
        throw new IllegalArgumentException("Column at position " + (i + 1) + " needs an AS alias");
    }
}

try {
    session.execute(ctasSql);
} catch (SemanticException e) {
    if (e.getCode() == SemanticErrorCode.COLUMN_NAME_NOT_SPECIFIED) {
        // position is in e.getMessage(); add aliases and retry
    } else throw e;
}

When it happens

Trigger: Statements whose output includes an unnamed expression column where a name is required: `CREATE TABLE t AS SELECT 1 + 1;`, `CREATE VIEW v AS SELECT count(1) ...` on engines requiring aliases, or queries whose descriptor fields lack names at the checked point.

Common situations: CTAS/view creation from SELECTs with computed columns without AS aliases; generated SQL missing aliases after refactoring; selecting constants or function results and using them as table columns.

Related errors

• MUST_BE_AGGREGATE_OR_GROUP_BY
• NESTED_AGGREGATION
• NESTED_WINDOW
• MUST_BE_AGGREGATION_FUNCTION
• NOT_SUPPORTED

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/12f1a85b5d8c5960. Report an issue: GitHub.