---
name: metabase_validate_card_payload
kind: function
lang: py
domain: infra
version: "1.0.0"
purity: pure
signature: "def metabase_validate_card_payload(payload: dict) -> list[str]"
description: "Valida la estructura de un payload de card de Metabase sin necesidad de red. Recorre todos los checks y acumula todos los issues en vez de abortar al primero. Retorna lista vacia si el payload es valido."
tags: [metabase, validation, card, pure, pre-flight, structural]
uses_functions: []
uses_types: []
returns: []
returns_optional: false
error_type: ""
imports: []
params:
  - name: payload
    desc: "dict con los campos de la card a validar: name, display, dataset_query, type (opcional), visualization_settings (opcional), parameters (opcional), archived (opcional)"
output: "lista de strings describiendo cada issue estructural encontrado; lista vacia indica payload valido listo para enviarse a POST/PUT /api/card"
tested: true
tests:
  - "card valido retorna lista vacia"
  - "card display invalido"
  - "card display ausente"
  - "card name ausente"
  - "card name vacio"
  - "card dataset query ausente"
  - "card dataset query sin database"
  - "card nativa sin sql"
  - "card nativa mbql5"
  - "card type invalido"
  - "card type valido"
  - "card visualization settings no dict"
  - "card parameters no list"
  - "card archived no bool"
  - "card acumula multiples errores"
test_file_path: "python/functions/metabase/validation_test.py"
file_path: "python/functions/metabase/validation.py"
---

## Ejemplo

```python
issues = metabase_validate_card_payload({
    "name": "Revenue by Month",
    "display": "line",
    "dataset_query": {
        "database": 1,
        "type": "native",
        "native": {"query": "SELECT date_trunc('month', created_at), SUM(total) FROM orders GROUP BY 1"},
    },
})
if issues:
    print("Payload invalido:", issues)
else:
    metabase_update_card(client, card_id, **payload)
```

## Notas

Displays validos: scalar, table, line, bar, pie, area, row, funnel, smartscalar, gauge, progress, combo, pivot, map, scatter, waterfall, sankey, object.

Tipos validos (campo `type`): question, model, metric.

Soporta dos formatos de SQL nativo:
- Legacy: `dataset_query.native.query`
- MBQL5: `dataset_query.stages[0].native`

No aborta al primer error — recolecta todos los issues para que el caller pueda mostrarlos todos de una vez.