[Qgis-user] QGIS and MSSQL schembound view

[Sergio Vieira Rolanski]

Hi,

Using QGIS 2.4 with SQLSERVER 2008 R2.
I'm creating the following schemabound view:

CREATE VIEW myQGisView
WITH SCHEMABINDING
AS
	SELECT  d.seq_roteiro * 1000 + d.seq_descarga AS qgis_id,
		d.seq_transmissao, d.seq_lote, d.seq_roteiro, d.seq_descarga, a.val_numero_quadra, a.val_face_quadra,
		b.seq_bairro, b.des_bairro,  l.des_logradouro, a.des_novo_logradouro, d.des_numero_imovel, d.des_complemento, d.des_cep,
		GEOGRAPHY::STGeomFromText('POINT(' + substring(a.des_coordenada_gps, charindex(',', a.des_coordenada_gps) + 1, 99) + ' ' + substring(a.des_coordenada_gps, 0, charindex(',', des_coordenada_gps)) +')', 4326) AS geo_localizacao
	FROM t1 d
		INNER JOIN t2 a ON d.seq_roteiro = a.seq_roteiro AND d.seq_descarga = a.seq_descarga
		INNER JOIN t3 l ON d.seq_logradouro = l.seq_logradouro
		INNER JOIN t4 b ON b.seq_bairro = d.seq_bairro
	WHERE d.seq_transmissao IS NOT NULL  
GO

CREATE UNIQUE CLUSTERED INDEX IDX_VW_CADASTRO_QGIS ON myQGisView(qgis_id)
GO

The view and indexes are created, the calculation I did fo qgis_id will be always unique.
QGIS has problems accepting it, error says "invalid layer and cannot be loaded."
and the dropbox for PK columns doesn't open up.
If I create a common view with this as first column in the select:

SELECT row_number() OVER(ORDER BY d.seq_lote, d.seq_transmissao, d.seq_roteiro, d.seq_descarga) AS qgis_id,
	
QGIS accept the view and everything works, but selecting this view is very, very slow.

Any ideias?

Thank you in advance,
Sergio