dbt-labs/dbt-core · error

type conversion from table column in MockAdapter

Error message

type conversion from table column in MockAdapter

What it means

AdapterImpl::convert_type panics when called on a MockAdapter. This method maps an Arrow field DataType from an AgateTable column into the adapter's SQL type string; the mock adapter has no meaningful type-mapping to offer, so the case is explicitly unimplemented.

Source

Thrown at crates/dbt-adapter/src/adapter/adapter_impl.rs:2357

            ) => {
                if quote_config.unwrap_or(true) {
                    Ok(self.quote(column))
                } else {
                    Ok(column.to_string())
                }
            }
        }
    }

    /// BaseAdapter https://github.com/dbt-labs/dbt-adapters/blob/0efd8d3d1081e1ab43e38797d5104f7b424a6284/dbt-adapters/src/dbt/adapters/base/impl.py#L1231
    pub fn convert_type(
        &self,
        state: &State,
        table: Arc<AgateTable>,
        col_idx: i64,
    ) -> AdapterResult<String> {
        if self.mock_state().is_some() {
            unimplemented!("type conversion from table column in MockAdapter")
        }
        let batch = table.original_record_batch();
        let schema = batch.schema();
        let data_type = schema.field(col_idx as usize).data_type();

        let data_type = match data_type {
            dt if dt.is_null() => &DataType::Int32,
            DataType::Float64 => {
                let is_int = batch
                    .column(col_idx as usize)
                    .as_any()
                    .downcast_ref::<arrow_array::Float64Array>()
                    .is_some_and(try_to_int_col);
                if is_int { &DataType::Int64 } else { data_type }
            }
            dt => dt,
        };

View on GitHub (pinned to 0267ce9170)

Solutions

1. Use a real adapter when type conversion of table columns is needed.
2. Avoid code paths that render column DDL/types from AgateTables in mock mode.
3. Extend the mock branch in convert_type to return a generic type (e.g. from the Arrow schema) instead of panicking.

Example fix

// before
if self.mock_state().is_some() {
    unimplemented!("type conversion from table column in MockAdapter")
}
// after
if self.mock_state().is_some() {
    return Ok(default_type_for(data_type));
}

if adapter.is_mock() {
    return Err("convert_type requires a real adapter".into());
}

fn supports_type_conversion(adapter: &AdapterImpl) -> bool { adapter.mock_state().is_none() }

When it happens

Trigger: Calling convert_type(state, table, col_idx) on an adapter in mock mode — typically during column schema rendering of loaded AgateTables under parse/mock runs.

Common situations: Seed or model column rendering pipelines running with a mock adapter; tests that load table data into AgateTable but never execute against a real adapter.

Understand the failure class

Background: Type mismatch errors: IllegalArgumentException, TypeError and type guards across 150 open-source libraries — this error's family across 150 libraries.

Related errors

• query addition to connection in MockAdapter
• load_dataframe() for the Salesforce adapter
• Starburst
• Athena
• Trino

AI-assisted analysis of dbt-labs/dbt-core@0267ce9170 (2026-09-07). Data as JSON: /api/errors/ed5e86c0a2a43c3a. Report an issue: GitHub.