conversion_type_params must exist for conversion metric

Error message

conversion_type_params must exist for conversion metric

What it means

Panic "conversion_type_params must exist for conversion metric" fires in resolve_top_level_metrics when a conversion metric's type_params.conversion_type_params is Some at the guard but the cloned/checked value is unexpectedly None at unwrap time — practically, this occurs when metric type is conversion but conversion_type_params was cleared or mis-constructed between the check and the expect.

Source

Thrown at crates/dbt-parser/src/resolve/resolve_metrics.rs:503

                    .cumulative_type_params
                    .clone()
                    .unwrap_or_default()
                    .metric;

                if let Some(metric) = maybe_cumulative_metric {
                    vec![metric.name]
                } else {
                    vec![]
                }
            }
            MetricType::Conversion => {
                if type_params.conversion_type_params.is_none() {
                    vec![]
                } else {
                    let conversion_type_params = type_params
                        .conversion_type_params
                        .clone()
                        .expect("conversion_type_params must exist for conversion metric");
                    let base_metric_name =
                        conversion_type_params.base_metric.unwrap_or_default().name;
                    let conversion_metric_name = conversion_type_params
                        .conversion_metric
                        .unwrap_or_default()
                        .name;

                    if base_metric_name.is_empty() || conversion_metric_name.is_empty() {
                        vec![]
                    } else {
                        vec![base_metric_name, conversion_metric_name]
                    }
                }
            }
            _ => vec![],
        };

        let unique_ids_of_nodes_depends_on: Vec<String> = names_of_nodes_depends_on

View on GitHub (pinned to 0267ce9170)

Solutions

Validate the conversion metric's YAML: ensure conversion_type_params with base_metric and conversion_metric are present.
Replace the guard+expect with a single match/if-let that produces a proper schema error naming the metric.
Check for any code path between the guard and the expect that resets conversion_type_params to None.

Example fix

// before
if type_params.conversion_type_params.is_none() { vec![] } else {
    let ctp = type_params.conversion_type_params.clone().expect("...");
// after
match type_params.conversion_type_params.clone() {
    Some(ctp) => { /* ... */ }
    None => return Err(err![arg.ctxt, "conversion metric '{}' missing conversion_type_params", name]),
}

Defensive patterns

Strategy: validation

Validate before calling

# validate conversion metric YAML before dbt parse
# metrics:
#   - name: conv
#     type: conversion
#     type_params:
#       conversion_type_params:
#         base_metric: {name: base}
#         conversion_metric: {name: conv_metric}
#         window: {count: 7, period: day}
assert 'conversion_type_params' in metric['type_params']

Prevention

Always define base_metric, conversion_metric, and window for conversion metrics
Run dbt parse (schema check) before deeper pipelines
Prefer match over is_none-check + expect for Option fields

When it happens

Trigger: Defining a metric with `type: conversion` whose conversion_type_params were dropped during deserialization (missing base_metric/conversion_metric fields) or mutated by config overrides before resolution.

Common situations: A conversion metric YAML missing required nested fields (window, base_metric, conversion_metric); schema changes in MetricTypeParams making the is_none guard stale.

Understand the failure class

Background: "is required", "must be set", "missing required field": configuration validation errors across open-source libraries — this error's family across 36 libraries.

Related errors

AI-assisted analysis of dbt-labs/dbt-core@0267ce9170 (2026-09-07). Data as JSON: /api/errors/4284ab9a15300ece. Report an issue: GitHub.