Use trigger to check geometry validity on first import

Instead of importing all data and then only checking the geometry when
clustering the data. Because there are more things happening in parallel
on the first import this makes the whole import slightly faster.

Author: joto

src/flex-table.cpp

@@ -176,6 +176,12 @@ void table_connection_t::start(bool append)
             table().has_geom_column() ? flex_table_t::table_type::interim
                                       : flex_table_t::table_type::permanent,
             table().full_name()));
+
+        if (table().has_geom_column() && table().geom_column().srid() != 4326) {
+            create_geom_check_trigger(m_db_connection.get(), table().schema(),
+                                      table().name(),
+                                      table().geom_column().name());
+        }
     }
 
     prepare();
@@ -195,6 +201,11 @@ void table_connection_t::stop(bool updateable, bool append)
     util::timer_t timer;
 
     if (table().has_geom_column()) {
+        if (table().geom_column().needs_isvalid()) {
+            drop_geom_check_trigger(m_db_connection.get(), table().schema(),
+                                    table().name());
+        }
+
         log_info("Clustering table '{}' by geometry...", table().name());
 
         // Notices about invalid geometries are expected and can be ignored
@@ -207,14 +218,6 @@ void table_connection_t::stop(bool updateable, bool append)
         std::string sql = "INSERT INTO {} SELECT * FROM {}"_format(
             table().full_tmp_name(), table().full_name());
 
-        if (table().geom_column().srid() != 4326) {
-            // libosmium assures validity of geometries in 4326.
-            // Transformation to another projection could make the geometry
-            // invalid. Therefore add a filter to drop those.
-            sql += " WHERE ST_IsValid(\"{}\")"_format(
-                table().geom_column().name());
-        }
-
         auto const postgis_version = get_postgis_version(*m_db_connection);
 
         sql += " ORDER BY ";

src/pgsql-helper.cpp

@@ -55,6 +55,18 @@ void create_geom_check_trigger(pg_conn_t *db_connection,
                         qualified_name(schema, table), func_name));
 }
 
+void drop_geom_check_trigger(pg_conn_t *db_connection,
+                             std::string const &schema,
+                             std::string const &table)
+{
+    std::string func_name = qualified_name(schema, table + "_osm2pgsql_valid");
+
+    db_connection->exec("DROP TRIGGER \"{}\" ON {};"_format(
+        table + "_osm2pgsql_valid", qualified_name(schema, table)));
+
+    db_connection->exec("DROP FUNCTION IF EXISTS {} ();"_format(func_name));
+}
+
 void analyze_table(pg_conn_t const &db_connection, std::string const &schema,
                    std::string const &name)
 {

src/pgsql-helper.hpp

@@ -31,6 +31,10 @@ void create_geom_check_trigger(pg_conn_t *db_connection,
                                std::string const &table,
                                std::string const &geom_column);
 
+void drop_geom_check_trigger(pg_conn_t *db_connection,
+                             std::string const &schema,
+                             std::string const &table);
+
 void analyze_table(pg_conn_t const &db_connection, std::string const &schema,
                    std::string const &name);
 

src/table.cpp

@@ -132,6 +132,11 @@ void table_t::start(std::string const &conninfo, std::string const &table_space)
 
         //create the table
         m_sql_conn->exec(sql);
+
+        if (m_srid != "4326") {
+            create_geom_check_trigger(m_sql_conn.get(), m_target->schema,
+                                      m_target->name, "way");
+        }
     }
 
     prepare();
@@ -185,6 +190,11 @@ void table_t::stop(bool updateable, bool enable_hstore_index,
     if (!m_append) {
         util::timer_t timer;
 
+        if (m_srid != "4326") {
+            drop_geom_check_trigger(m_sql_conn.get(), m_target->schema,
+                                    m_target->name);
+        }
+
         log_info("Clustering table '{}' by geometry...", m_target->name);
 
         // Notices about invalid geometries are expected and can be ignored
@@ -195,13 +205,6 @@ void table_t::stop(bool updateable, bool enable_hstore_index,
             "CREATE TABLE {} {} AS SELECT * FROM {}"_format(
                 qual_tmp_name, m_table_space, qual_name);
 
-        if (m_srid != "4326") {
-            // libosmium assures validity of geometries in 4326.
-            // Transformation to another projection could make the geometry
-            // invalid. Therefore add a filter to drop those.
-            sql += " WHERE ST_IsValid(way)";
-        }
-
         auto const postgis_version = get_postgis_version(*m_sql_conn);
 
         sql += " ORDER BY ";