colouring-montreal/etl/load_new_geometries.sh

56 lines
1.8 KiB
Bash
Raw Normal View History

#!/usr/bin/env bash
2022-04-08 07:17:53 -04:00
# Load geometries from GeoJSON to Postgres
# - assume postgres connection details are set in the environment using PGUSER, PGHOST etc.
: ${1?"Usage: $0 ./path/to/mastermap/dir"}
mastermap_dir=$1
# Create 'geometry' record with
# id: <polygon-guid>,
# source_id: <toid>,
# geom: <geom>
2022-04-14 04:55:58 -04:00
echo "Removing temp tables if previously created..."
psql -c "DROP TABLE IF EXISTS new_geometries;"
psql -c "DROP TABLE IF EXISTS release_geometries;"
echo "Creating temporary geometries table for OS release geometries..."
psql -c "CREATE TABLE IF NOT EXISTS release_geometries (
geometry_id serial,
source_id varchar(30),
geometry_geom geometry(GEOMETRY, 3857)
);"
echo "Copy geometries to db..."
find $mastermap_dir -type f -name '*.3857.csv' \
-printf "$mastermap_dir/%f\n" | \
parallel \
cat {} '|' psql -c "\"COPY release_geometries ( geometry_geom, source_id ) FROM stdin WITH CSV HEADER;\""
2022-04-14 04:55:58 -04:00
echo "Creating temporary geometries table for new geometries only..."
psql -c "CREATE TABLE IF NOT EXISTS new_geometries (
source_id varchar(30),
2022-04-14 04:55:58 -04:00
geometry_geom geometry(GEOMETRY, 3857)
);"
# Delete any duplicated geometries (by TOID)
echo "Delete duplicate geometries..."
psql -c "DELETE FROM release_geometries a USING (
SELECT MIN(ctid) as ctid, source_id
FROM release_geometries
GROUP BY source_id
HAVING COUNT(*) > 1
) b
WHERE a.source_id = b.source_id
AND a.ctid <> b.ctid;"
2022-04-14 04:55:58 -04:00
echo "Finding geometries that are new to this release..."
psql -c "INSERT INTO new_geometries ( source_id, geometry_geom )
SELECT source_id, geometry_geom
2022-04-14 05:33:38 -04:00
FROM release_geometries AS r
WHERE NOT EXISTS ( SELECT source_id
FROM geometries AS g
WHERE g.source_id = r.source_id);"