Database: make conversion & helmert_transformation updatable views

- Transform conversion as a view, and when inserting into it, actually
  insert into 3 tables: conversion_table, conversion_method and conversion_param,
  so that method and parameter names are not repeated each time.
- Similarly for helmert_tranformation, insert into helmert_transformation_tabl
  and coordinate_operation_method.

This reduces the db size from 6 344 704 bytes to 5 853 184 bytes, without
significant slowdown for queries.

1 files changed, 1 insertions, 1 deletions

diff --git a/data/sql/commit.sql b/data/sql/commit.sql
index 7aa91e8e..eb49828a 100644
--- a/data/sql/commit.sql
+++ b/data/sql/commit.sql
@@ -4,7 +4,7 @@ CREATE INDEX geodetic_crs_datum_idx ON geodetic_crs(datum_auth_name, datum_code)
 CREATE INDEX geodetic_datum_ellipsoid_idx ON geodetic_datum(ellipsoid_auth_name, ellipsoid_code);
 CREATE INDEX supersession_idx ON supersession(superseded_table_name, superseded_auth_name, superseded_code);
 CREATE INDEX deprecation_idx ON deprecation(table_name, deprecated_auth_name, deprecated_code);
-CREATE INDEX helmert_transformation_idx ON helmert_transformation(source_crs_auth_name, source_crs_code, target_crs_auth_name, target_crs_code);
+CREATE INDEX helmert_transformation_idx ON helmert_transformation_table(source_crs_auth_name, source_crs_code, target_crs_auth_name, target_crs_code);
 CREATE INDEX grid_transformation_idx ON grid_transformation(source_crs_auth_name, source_crs_code, target_crs_auth_name, target_crs_code);
 CREATE INDEX other_transformation_idx ON other_transformation(source_crs_auth_name, source_crs_code, target_crs_auth_name, target_crs_code);
 CREATE INDEX concatenated_operation_idx ON concatenated_operation(source_crs_auth_name, source_crs_code, target_crs_auth_name, target_crs_code);