back to first page [..] [[BR]]
back to ["CookBook Eda"] [[BR]]
back to ["Obstacle pressure"] [[BR]]

= Join ROE - CCM =
== Putting geobs2010 into the CCM ==
Initially, the database was located in another table, below we just copy the table shema roe into the eda2.0 database

I have changed the name geobs into geobs2010 with the new version of the ROE.
{{{
REM pour Cédric
CD C:\Documents and Settings\cedric\Mes documents\Migrateur\eda\Barrages
C:\"Program Files"\PostgreSQL\8.4\bin\pg_dump -U postgres -p 5433 -n geobs2010 ROE> roe_schema.sql
C:\"Program Files"\PostgreSQL\8.4\bin\psql -d eda2.0 -h localhost -p 5433 -U postgres -f roe_schema.sql
REM====================================
REM Pour Céline
CD D:\Celine Jouanin\ROE
C:\"Program Files"\PostgreSQL\8.4\bin\pg_dump -U postgres  -n geobs2010 ROE> roe_schema.sql
C:\"Program Files"\PostgreSQL\8.4\bin\psql -d eda2.0 -h localhost -p 5432 -U postgres -f roe_schema.sql
}}}


== We have to reproject our data ==
The constraint CONSTRAINT enforce_srid_ref_position_locale CHECK (srid(ref_position_locale) = 27572)
will make it impossible to change the srid from the roe (is that right ?... yes)
We will create a new column a new constraint and fill this column with reprojected data
{{{
select AddGeometryColumn('geobs2010', 'obstacle_referentiel','ref_position_etrs89' , 3035,'POINT',2);
-- this function also creates constraint « enforce_srid_ref_position_etrs89 
select SRID (ref_position_locale) FROM geobs2010.obstacle_referentiel; -- all lines 27572 (Lambert II)
update geobs2010.obstacle_referentiel set ref_position_etrs89 =ST_Transform(ref_position_locale,3035);
}}}


== Select only the dams within France ==
=== Extract the data within the french boundaries ===
this is ticket #45

==== With !ArcGis ====
Use the french limits (see ["French boundaries"]) in RGF93[[BR]]
In __!QuantumGis__, connect the ROE database and add the obstacle_referentiel layer[[BR]]
Clic right on the layer "Sauvegarder comme shapefile" with Fichiers de type : Shapefiles (*.shp), Codage=UTF-8 > ok
and choose "Système de coordonnées de référence" = RGF93/Lambert93 (because the french boundary layer is in this system).[[BR]]
In __!ArcGis__, add the two layers :[[BR]]
- ROE_EDA.shape[[BR]]
- GEOFLADept_FR_Corse_AV_L93\DEPARTEMENT.shp[[BR]]
Use Analysis Tools > Extract > Clip[[BR]]
Input Features : ROE_EDA[[BR]]
Clip Features : DEPARTEMENT[[BR]]

If you want you can choose a "XY Tolerance", because some points in the ROE are in France but they are at the limit.
(maybe 300meters?).[[BR]]

==== With QGis ====

first change projection to ERTRS 89[[BR]]
previously we have done this with a request in postgis, this is how you can use gdal
look at ["Cookbook gdal"] =>paragraph "changing the projection of any source"[[BR]]
write 
{{{
ogr2ogr -f formatdst_datasource_name src_datasource_name -a_srs <actual projection> -t_srs <reprojected data>
}}}
write ogr2ogr -f formatdst_datasource_name src_datasource_name -a_srs <actual projection> -t_srs <reprojected data>[[BR]]
Launch in ''' OSGeo4W command shell '''
{{{
CD C:\Documents and Settings\cedric\Mes documents\Migrateur\eda\couches_SIG\france
C:\OSGeo4W\bin\ogr2ogr DEPARTEMENT_3035.shp DEPARTEMENT.shp -s_srs "EPSG:2154" -t_srs "EPSG:3035"
}}}
The shape file DEPARTEMENT_3035.shp is created with right projection[[BR]]
Connect the ROE database and add this newly projected obstacle_referential layer [[BR]]
Process 
Add the file GEOFLADept_FR_Corse_AV_L93\DEPARTEMENT.shp (layer>add a vector layer) [[BR]]
Tools > Geoprocessing tools > Clip[[BR]]
But this does not allow to select shapes within a distance, we put this base in postgis, here we create shape along with an index[[BR]]
{{{
REM Cédric =======================
CD C:\Documents and Settings\cedric\Mes documents\Migrateur\eda\couches_SIG\france
C:\"Program Files"\PostgreSQL\8.4\bin\shp2pgsql -s 3035 -I DEPARTEMENT_3035 france.departement> departement.sql
C:\"Program Files"\PostgreSQL\8.4\bin\psql -d eda2.0 -h localhost -U postgres -p 5433 -c "DROP schema if exists france CASCADE"
C:\"Program Files"\PostgreSQL\8.4\bin\psql -d eda2.0 -h localhost -U postgres -p 5433 -c "CREATE schema france "
C:\"Program Files"\PostgreSQL\8.4\bin\psql -d eda2.0 -h localhost -U postgres -p 5433 -f departement.sql
REM Céline =======================
CD D:\Celine Jouanin\france
C:\"Program Files"\PostgreSQL\8.4\bin\shp2pgsql -s 3035 -I DEPARTEMENT_3035 france.departement> departement.sql
C:\"Program Files"\PostgreSQL\8.4\bin\psql -d eda2.0 -h localhost -U postgres -p 5432 -c "DROP schema if exists france"
C:\"Program Files"\PostgreSQL\8.4\bin\psql -d eda2.0 -h localhost -U postgres -p 5432 -c "CREATE schema france"
C:\"Program Files"\PostgreSQL\8.4\bin\psql -d eda2.0 -h localhost -U postgres -p 5432 -f departement.sql
}}}
insertion dans la table geometry column
{{{
INSERT INTO geometry_columns(f_table_catalog, f_table_schema, f_table_name, f_geometry_column, coord_dimension, srid, "type")
SELECT '', 'france', 'departement', 'the_geom', ST_CoordDim(the_geom), ST_SRID(the_geom), GeometryType(the_geom)
FROM france.departement LIMIT 1;
}}}
mmh so finally I would like a new column saying whether or not the dams are in France [[BR]]
 we have 43937 lines, after joining with France we should have less
{{{
SELECT count(ref_id) from geobs.obstacle_referentiel -- 43937
}}}
{{{
SELECT count(ref_id) from geobs2010.obstacle_referentiel -- 47078
}}}


The following does not work and it makes sense, we should only have one geom of france
{{{
SELECT  count(ref_id) from geobs.obstacle_referentiel o
 join france.departement f ON ST_DWithin(f.the_geom, o.ref_position_etrs89, 2000); --49468
}}}
{{{
SELECT  count(ref_id) from geobs2010.obstacle_referentiel o
 join france.departement f ON ST_DWithin(f.the_geom, o.ref_position_etrs89, 2000); --52986
}}}
I first aggregate all departements and then join this glued layer with the dam layer, the distance is 2000
{{{
SELECT  count(ref_id) from geobs2010.obstacle_referentiel o
join (SELECT ST_Union(f.the_geom) as singlegeom
	 FROM france.departement  As f) as sub
ON ST_DWithin(sub.singlegeom, o.ref_position_etrs89, 2000); --47049 
}}}

[[Image(source:data/Docs/trac/barrages_cutfrance.png,600px)]][[BR]]


== Build a view allowing to know the score of the different dams ==
this is ticket #46


== Joining CCM and Dams ==

-- Création d'un index sur obstacle_referentiel

-- Création de la table correspondance (roe_ccm_300), je rajoute un the_geom a la requete pour pouvoir voir la table dans Qgis 
'''La distance de projection entre la CCM et la base geobs est choisie à : 300 m (comme pour l'extraction des données BDMAP sur la CCM)''' [[BR]]
Attention à l'ordre des colonnes dans la clause group by, si un predicat min est utilisé il faut qu'il apparaisse en deuxième dans la clause [[BR]]
The table should contain
 * gid
 * dam_height
 * dam_score
{{{
CREATE INDEX indexgeobs ON geobs2010.obstacle_referentiel  USING GIST (ref_position_etrs89 GIST_GEOMETRY_OPS);
DROP TABLE IF EXISTS geobs2010.roe_ccm_300;
CREATE TABLE geobs2010.roe_ccm_300 as (
        SELECT distinct on (ref_id) ref_id, gid, min(distance) as distance, ref_position_etrs89 as the_geom FROM (
               SELECT ref_id, gid ,CAST(distance(r.the_geom, b.ref_position_etrs89) as  decimal(15,1)) as distance ,b.ref_position_etrs89
               FROM geobs2010.obstacle_referentiel As b
               INNER JOIN  ccm21.riversegments r ON ST_DWithin(r.the_geom, b.ref_position_etrs89,300)
--             WHERE b.goodproj IS TRUE -- Attention c'est faux
               ORDER BY ref_id) AS sub 
        GROUP BY ref_id, distance, gid, the_geom
); --53s


-- mise à jour de la table geometry_columns
select Probe_Geometry_Columns(); 
-- si la commande ci dessus ne marche pas ...
INSERT INTO geometry_columns(f_table_catalog, f_table_schema, f_table_name, f_geometry_column, coord_dimension, srid, "type")
SELECT '', 'geobs2010', 'roe_ccm_300', 'the_geom', ST_CoordDim(the_geom), ST_SRID(the_geom), GeometryType(the_geom)
FROM geobs2010.roe_ccm_300 LIMIT 1;

-- creation d'index, clé primaire, et constraintes qui vont bien
alter table geobs2010.roe_ccm_300 add column id serial PRIMARY KEY;
alter table geobs2010.roe_ccm_300 add CONSTRAINT enforce_dims_the_geom CHECK (ndims(the_geom) = 2);
alter table geobs2010.roe_ccm_300 add  CONSTRAINT enforce_geotype_the_geom CHECK (geometrytype(the_geom) = 'POINT'::text OR the_geom IS NULL);
alter table geobs2010.roe_ccm_300 add  CONSTRAINT enforce_srid_the_geom CHECK (srid(the_geom) = 3035);

CREATE INDEX indexgeobsroe_ccm_300 ON geobs2010.roe_ccm_300
  USING GIST ( the_geom GIST_GEOMETRY_OPS );
}}}
The final script is below
{{{
ALTER TABLE geobs2010.obstacle_referentiel ADD COLUMN goodproj boolean;
UPDATE geobs2010.obstacle_referentiel set goodproj=FALSE ;
UPDATE geobs2010.obstacle_referentiel set goodproj=TRUE where ref_id IN (
SELECT  ref_id from geobs2010.roe_ccm_300);
}}}
We now have to use, and can display the dams left out in qgis
{{{
select * from geobs2010.obstacle_referentiel where goodproj=TRUE
}}}


== Some statistics about the dam projection ==
We create a table of regions
{{{
DROP TABLE if exists france.region;
CREATE TABLE france.region as(
SELECT ST_Union(the_geom) as the_geom,code_reg FROM france.departement GROUP BY code_reg);
SELECT geometrytype(the_geom) from france.region;
Alter table france.region ADD column reg_id serial PRIMARY KEY;
alter table france.region add CONSTRAINT enforce_dims_the_geom CHECK (ndims(the_geom) = 2);
alter table france.region add  CONSTRAINT enforce_geomtype_the_geom CHECK (geometrytype(the_geom) = 'MULTIPOLYGON'::text OR geometrytype(the_geom) = 'POLYGON'::text OR the_geom IS NULL);
alter table france.region add  CONSTRAINT enforce_srid_the_geom CHECK (srid(the_geom) = 3035);
}}}
Calcul du pourcentage de barrages projetés par région.
{{{
DROP TABLE if exists geobs2010.pourcentageprojette;
CREATE TABLE geobs2010.pourcentageprojette AS(
SELECT 100*JOINT/(JOINT +DISJOINT) as pourcentage, nom_region,the_geom FROM( 
SELECT nom_region, sum(cast(cast(goodproj AS integer)as NUMERIC)) as joint, sum(cast(cast(NOT goodproj AS integer) AS NUMERIC)) as disjoint, the_geom
from  france.region f, geobs2010.obstacle_referentiel g
WHERE ST_Intersects(f.the_geom,g.ref_position_etrs89)
group by nom_region, the_geom)as sub);
SELECT geometrytype(the_geom) FROM geobs2010.pourcentageprojette;
ALTER TABLE geobs2010.pourcentageprojette ADD column pourc_id serial PRIMARY KEY;
alter table geobs2010.pourcentageprojette  add CONSTRAINT enforce_dims_the_geom CHECK (ndims(the_geom) = 2);
alter table geobs2010.pourcentageprojette  add  CONSTRAINT enforce_geomtype_the_geom CHECK(geometrytype(the_geom) = 'MULTIPOLYGON'::text OR geometrytype(the_geom) = 'POLYGON'::text OR the_geom IS NULL);;
alter table geobs2010.pourcentageprojette  add  CONSTRAINT enforce_srid_the_geom CHECK (srid(the_geom) = 3035);
}}}
Pourcentage de barrages projettés avec un buffer de 300 m

[[Image(source:data/Docs/trac/pourc_projette.jpg)]]

(Avec un buffer de 1000, 2000m j'ai toujours les mêmes pourcentages pour chaque région, soit c'est normal soit la requête n'a pas fonctionné).[[BR]]


When we compare the geobs2010.roe_ccm_300 and the geobs2010.obstacle_referentiel :[[BR]]
[[Image(source:data/Docs/trac/ROE/Obstacle referentiel et correspondance.png,800px)]][[BR]]
in red the dams in geobs2010.obstacle_referentiel which are not well projected
in violet the dams which are "well" projected and projected in the CCM layer[[BR]]
With geobs2010.roe_ccm_300, there are two dams (in violet in the East) which are projected in the CCM layer but aren't in France :[[BR]]
ref_id =21164  an ref_id=49967