Feat: improve performances for syntheseff view

[jpm-cbna]

See #390.

[jpm-cbna]

Après discussion avec @TheoLechemia , il serait plus pertinent de basculer cette VM dans le schéma synthese car elle ne devrait être utilisé que par la vue synthese.syntheseff.

[amandine-sahl]

Je me demande si pour des questions de performance, et etant donnée que ce sont des données "stables", il ne serait pas plus simple d'avoir une table de précalcul de la correspondance entre sensibilité, niveau de diffusion. Que l'on pourrait ensuite utiliser dans l'atlas.

CREATE TABLE gn_sensitivity.cor_diffusion_level_sensitivity AS
WITH sens AS (
    SELECT  id_nomenclature AS id_sens, cd_nomenclature  AS cd_sens,
    CASE
          WHEN cd_nomenclature = '0' THEN '5'
          ELSE cd_nomenclature 
      END AS cor_cd_diff,
        cd_nomenclature::int AS order_sens
    FROM ref_nomenclatures.t_nomenclatures AS tn
    WHERE id_type =  ref_nomenclatures.get_id_nomenclature_type('SENSIBILITE')
    UNION
    SELECT -1 AS id_sens, NULL  AS cd_sens, NULL AS cor_cd_diff, 0  AS order_sens
) , diff AS (
    SELECT id_nomenclature AS id_diff, cd_nomenclature  AS cd_diff,
     CASE
          WHEN cd_nomenclature IN ('0', '5') THEN 0
          ELSE cd_nomenclature::int
        END   AS order_diff
    FROM ref_nomenclatures.t_nomenclatures AS tn
    WHERE id_type =  ref_nomenclatures.get_id_nomenclature_type('NIV_PRECIS')
    UNION
    SELECT -1 AS id_diff, NULL  AS cd_diff, 0  AS order_diff
), d AS (
    SELECT * , GREATEST(sens.order_sens, diff.order_diff) AS order_max,
     CASE
          WHEN GREATEST(sens.order_sens, diff.order_diff) = 0 THEN '5'
          ELSE GREATEST(sens.order_sens, diff.order_diff)::varchar
        END   AS cd_max
    FROM sens, diff
)
SELECT d.*,  tn.id_nomenclature AS id_max
FROM d
JOIN ref_nomenclatures.t_nomenclatures AS tn
ON id_type =  ref_nomenclatures.get_id_nomenclature_type('NIV_PRECIS') AND cd_nomenclature = cd_max;

SELECT DISTINCT s.id_nomenclature_sensitivity , s.id_nomenclature_diffusion_level , cds.*
FROM gn_synthese.synthese AS s 
JOIN gn_sensitivity.cor_diffusion_level_sensitivity cds
ON cds.id_sens = COALESCE(s.id_nomenclature_sensitivity, -1) AND cds.id_diff = COALESCE(s.id_nomenclature_diffusion_level, -1)