--- pg_extension_name: pg_xenophile pg_extension_version: 0.7.4 pg_readme_generated_at: 2023-05-29 12:18:20.020209+01 pg_readme_version: 0.6.4 --- # `pg_xenophile` PostgreSQL extension [![PGXN version](https://badge.fury.io/pg/pg_xenophile.svg)](https://badge.fury.io/pg/pg_xenophile) The `pg_xenophile` PostgreSQL extension bundles a bunch of data, data structures and routines that you often end up needing when working on an international project: - tables with the usual data that you need on countries, regions, languages and currencies; - functions to easily store and access translated strings; and - (trigger) functions to set up one-to-many translation tables with easy-to-use views on top. It's perfectly valid to _just_ use `pg_xenophile` as a repository for up-to-date lists of countries and languages and such. But, the extension becomes especially worthwhile if you want some comfort on top of the common many-to-one translation-table pattern. > All your ethnocentrism are belong to us. ## Using `pg_xenophile` To use the list of countries (from the [`country` table](#table-country)) or languages (from the [`lang` table](#table-lang), just use them. And don't be afraid of using the natural keys in your foreign keys! If you've been told to fear them, you will soon be attached to the convenience of not needing to join to know what a foreign key value means. If you want a translatable table, you have to register the base table with the [`l10n_table` meta table](#table-l10n_table). See the [`l10n_table` documentation](#table-l10n_table) in the reference for details. From the `l10n_table` documentation, you should also be able to learn how to work with the `lang_l10n`, `lang_l10n_en`, `country_l10n` and `country_l10n_en` tables and views that are manintained via the triggers on this meta table. ## Extension-specific settings | Extenion-hooked setting name | `app.`-hooked setting name | Default setting value | | -------------------------------- | -------------------------------------- | ------------------------------- | | `pg_xenophile.base_lang_code` | `app.settings.i18n.base_lang_code` | `'en'::xeno.lang_code_alpha2` | | `pg_xenophile.user_lang_code` | `app.settings.i18n.user_lang_code` | `'en'::xeno.lang_code_alpha2` | | `pg_xenophile.target_lang_codes` | `app.settings.i18n.target_lang_codes` | `'{}'::xeno.lang_code_alpha2[]` | The reason that each `pg_xenophile` setting has an equivalent setting with an `app.settings.i18n` prefix is because the powerful PostgREST can pass on such settings from environment variables: `PGRST_APP_SETTINGS_*` maps to `app.settings.*`. The `app.settings.`-prefixed settings take precedence over `pg_xenophile.`-prefixed settings. Supporting _only_ the `app.settings.`-prefixed settings would not be a good idea, because, in the circumstance that you would be running an extension called “`app`”, these settings might disappear, as per the [relevant documentation](https://www.postgresql.org/docs/15/runtime-config-custom.html): > […] Such variables are treated as placeholders and have no function until > the module that defines them is loaded. When an extension module is loaded, it > will add its variable definitions and convert any placeholder values according > to those definitions. If there are any unrecognized placeholders that begin > with its extension name, warnings are issued and those placeholders are > removed. In addition to the above, the `user_lang_code` setting, if set as neither `app.settings.i18n.user_lang_code` and `pg_xenophile.user_lang_code`, falls back to the first two letters of the `lc_messages` setting. ### Internal settings | Setting name | Default setting value | | -------------------------------------------- | ------------------------------- | | `pg_xenophile.in_l10n_table_event_trigger` | `false` | | `pg_xenophile.in_l10n_table_row_trigger` | `false` | | `pg_xenophile.pg_restore_seems_active` | `false` | ## Object reference ### Schema: `xeno` `pg_xenophile` must be installed in the `xeno` schema. Hence, it is not relocatable. --- The `xeno` schema belongs to the `pg_xenophile` extension. Postgres (as of Pg 15) doesn't allow one to specify a _default_ schema, and do something like `schema = 'xeno'` combined with `relocatable = true` in the `.control` file. Therefore I decided to bluntly force the `xeno` schema name upon you, even though you might have very well (and justifyingly so) preferred something like `i18n`. ### Tables There are 8 tables that directly belong to the `pg_xenophile` extension. #### Table: `l10n_table` The `l10n_table` table is meant to keep track and manage all the `_l10n`-suffixed tables. By inserting a row in this table, with just the details of the base table, a many-to-one l10n table called `_l10n` will be created by the `maintain_l10n_objects` trigger. This trigger will also take care of creating the `_l10n_` view as well as one such view for all the `target_lang_codes`. These views combine the columns of the base table with the columns of the l10n table, filtered by the language code specific to that particular view. One of the reasons to manage this through a table rather than through a stored procedure is that a list of such enhanced l10n tables needs to be kept by `pg_xenophile` anyway: in the likely case that updates necessitate the upgrading of (the views and/or triggers around) these tables, the extension update script will know where to find everything. It may not immediately be obvious why, besides the `base_table_regclass` and the `l10n_table_regclass` columns, `schema_name`, `base_table_name` and `l10n_table_name` also exist. After all, PostgreSQL has some very comfortable magic surrounding `regclass` and related [object identifier types](https://www.postgresql.org/docs/current/datatype-oid.html). Two reasons: 1. Even though `pg_dump` has the ability to dump OIDs, tables belonging to extensions are not dumped at all, except for any part exempted from this using the `pg_catalog.pg_extension_config_dump()` function. For `l10n_table`, only the columns for which `l10n_table_belongs_to_extension_name IS NULL` are included in the dump. 2. OIDs of tables and other catalog objects are not guaranteed to remain the same between `pg_dump` and `pg_restore`. The `l10n_table` table has 12 attributes: 1. `l10n_table.schema_name` `name` - `NOT NULL` - `DEFAULT CURRENT_SCHEMA` 2. `l10n_table.base_table_name` `name` - `NOT NULL` 3. `l10n_table.base_table_regclass` `regclass` - `NOT NULL` - `PRIMARY KEY (base_table_regclass)` 4. `l10n_table.base_column_definitions` `text[]` - `NOT NULL` 5. `l10n_table.l10n_table_name` `name` - `NOT NULL` 6. `l10n_table.l10n_table_regclass` `regclass` - `NOT NULL` - `UNIQUE (l10n_table_regclass)` 7. `l10n_table.l10n_column_definitions` `text[]` - `NOT NULL` 8. `l10n_table.l10n_table_constraint_definitions` `text[]` - `NOT NULL` - `DEFAULT ARRAY[]::text[]` 9. `l10n_table.base_lang_code` `lang_code_alpha2` - `NOT NULL` - `DEFAULT pg_xenophile_base_lang_code()` 10. `l10n_table.target_lang_codes` `lang_code_alpha2[]` - `NOT NULL` - `DEFAULT pg_xenophile_target_lang_codes()` 11. `l10n_table.l10n_table_belongs_to_extension_name` `name` This column must be `NOT NULL` if the l10n table is created through extension setup scripts and its row in the meta table must thus be omitted from `pg_dump`. If `l10n_table_belongs_to_extension_name IS NOT NULL`, then the created localization (l10n) _table_ will be managed (and thus recreated during a restore) by the named extension's setup/upgrade script. That is _not_ the same as saying that the l10n table's _rows_ will belong to `pg_xenophile`. To determine the latter, a `l10n_columns_belong_to_extension_name` column will be added to the l10n table if the `l10n_table__maintain_l10n_objects()` trigger function finds `l10n_table_belongs_to_extension_name IS NOT NULL` on insert. Only developers of this or dependent extensions need to worry about these booleans. For users, the default of `false` assures that they will lose none of their precious data. 12. `l10n_table.l10n_table_belongs_to_extension_version` `text` #### Table: `lang` The `lang` table has 2 attributes: 1. `lang.lang_code` `lang_code_alpha2` ISO 639-1 two-letter (lowercase) language code. - `NOT NULL` - `PRIMARY KEY (lang_code)` 2. `lang.lang_belongs_to_pg_xenophile` `boolean` `pg_dump` will ignore rows for which this is `true`. Make sure that this column is `false` when you add your own language. When your language is an official language according to the ISO standard, please make sure that it will be included upstream in `pg_xenophile`, so that all users of the extension can profit from up-to-date information. Please note, that you will run into problems with dump/restore when you add records to this table from within your own dependent extension set up scripts. - `NOT NULL` - `DEFAULT false` #### Table: `lang_l10n` This table is managed by the `pg_xenophile` extension, which has delegated its creation to the `maintain_l10n_objects` trigger on the `l10n_table` table. To alter this table, just `ALTER` it as you normally would. The `l10n_table__track_alter_table_events` event trigger will detect such changes, as well as changes to the base table (`lang`) referenced by the foreign key (that doubles as primary key) on `lang_l10n`. When any `ALTER TABLE lang_l10n` or `ALTER TABLE lang` events are detected, `l10n_table` will be updated—the `base_column_definitions`, `l10n_column_definitions` and `l10n_table_constraint_definitions` columns—with the latest information from the `pg_catalog`. These changes to `l10n_table` in turn trigger the `maintain_l10n_objects` trigger, which ensures that the language-specific convenience views that (left) join `lang` to `lang_l10n` are kept up-to-date with the columns in these tables. To drop this table, either just `DROP TABLE` it (and the `l10n_table__track_drop_table_events` will take care of the book-keeping or delete its bookkeeping row from `l10n_table`. The `lang_l10n` table has 5 attributes: 1. `lang_l10n.lang_code` `lang_code_alpha2` - `NOT NULL` - `FOREIGN KEY (lang_code) REFERENCES lang(lang_code) ON UPDATE CASCADE ON DELETE CASCADE` 2. `lang_l10n.l10n_lang_code` `lang_code_alpha2` - `NOT NULL` - `FOREIGN KEY (l10n_lang_code) REFERENCES lang(lang_code) ON UPDATE RESTRICT ON DELETE RESTRICT` 3. `lang_l10n.l10n_columns_belong_to_extension_name` `name` 4. `lang_l10n.l10n_columns_belong_to_extension_version` `text` 5. `lang_l10n.name` `text` - `NOT NULL` #### Table: `country_l10n` This table is managed by the `pg_xenophile` extension, which has delegated its creation to the `maintain_l10n_objects` trigger on the `l10n_table` table. To alter this table, just `ALTER` it as you normally would. The `l10n_table__track_alter_table_events` event trigger will detect such changes, as well as changes to the base table (`country`) referenced by the foreign key (that doubles as primary key) on `country_l10n`. When any `ALTER TABLE country_l10n` or `ALTER TABLE country` events are detected, `l10n_table` will be updated—the `base_column_definitions`, `l10n_column_definitions` and `l10n_table_constraint_definitions` columns—with the latest information from the `pg_catalog`. These changes to `l10n_table` in turn trigger the `maintain_l10n_objects` trigger, which ensures that the language-specific convenience views that (left) join `country` to `country_l10n` are kept up-to-date with the columns in these tables. To drop this table, either just `DROP TABLE` it (and the `l10n_table__track_drop_table_events` will take care of the book-keeping or delete its bookkeeping row from `l10n_table`. The `country_l10n` table has 5 attributes: 1. `country_l10n.country_code` `country_code_alpha2` - `NOT NULL` - `FOREIGN KEY (country_code) REFERENCES country(country_code) ON UPDATE CASCADE ON DELETE CASCADE` 2. `country_l10n.l10n_lang_code` `lang_code_alpha2` - `NOT NULL` - `FOREIGN KEY (l10n_lang_code) REFERENCES lang(lang_code) ON UPDATE RESTRICT ON DELETE RESTRICT` 3. `country_l10n.l10n_columns_belong_to_extension_name` `name` 4. `country_l10n.l10n_columns_belong_to_extension_version` `text` 5. `country_l10n.name` `text` - `NOT NULL` #### Table: `currency` The `currency` table contains the currencies known to `pg_xenophile`. The `currency` table has 5 attributes: 1. `currency.currency_code` `currency_code` `currency_code` is a 3-letter ISO 4217 currency code. - `NOT NULL` - `PRIMARY KEY (currency_code)` 2. `currency.currency_code_num` `text` `currency_code` is the numeric 3-digit ISO 4217 currency code. - `NOT NULL` - `CHECK (currency_code_num ~ '^[0-9]{3}$'::text)` - `UNIQUE (currency_code_num)` 3. `currency.currency_symbol` `text` - `NOT NULL` - `CHECK (length(currency_symbol) = 1)` 4. `currency.decimal_digits` `integer` - `NOT NULL` - `DEFAULT 2` 5. `currency.currency_belongs_to_pg_xenophile` `boolean` Does this currency belong to the `pg_xenophile` extension or not. If `NOT currency_belongs_to_pg_xenophile`, it is considered a custom currency inserted by the extension user rather than the extension developer. Instead (or in addition) of adding such custom rows, please feel free to submit patches with all the currencies that you wish for `pg_xenophile` to embrace. - `NOT NULL` - `DEFAULT false` #### Table: `country` The ISO 3166-1 alpha-2, alpha3 and numeric country codes, as well as some auxillary information. The `country` table has 6 attributes: 1. `country.country_code` `country_code_alpha2` - `NOT NULL` - `PRIMARY KEY (country_code)` 2. `country.country_code_alpha3` `text` - `CHECK (country_code_alpha3 ~ '^[A-Z]{3}$'::text)` - `UNIQUE (country_code_alpha3)` 3. `country.country_code_num` `text` - `NOT NULL` - `CHECK (country_code_num ~ '^[0-9]{3}$'::text)` 4. `country.calling_code` `integer` - `NOT NULL` 5. `country.currency_code` `text` - `NOT NULL` - `DEFAULT 'EUR'::text` - `FOREIGN KEY (currency_code) REFERENCES currency(currency_code) ON UPDATE CASCADE ON DELETE RESTRICT` 6. `country.country_belongs_to_pg_xenophile` `boolean` `pg_dump` will ignore rows for which this is `true`. Make sure that this column is `false` when you add your own country. When your country is an official country according to the ISO standard, please make sure that it will be included upstream in `pg_xenophile`, so that all users of the extension can profit from up-to-date information. Please note, that you will run into problems with dump/restore when you add records to this table from within your own dependent extension set up scripts. - `NOT NULL` - `DEFAULT false` #### Table: `country_postal_code_pattern` The `country_postal_code_pattern` table has 8 attributes: 1. `country_postal_code_pattern.country_code` `country_code_alpha2` - `NOT NULL` - `PRIMARY KEY (country_code)` - `FOREIGN KEY (country_code) REFERENCES country(country_code)` 2. `country_postal_code_pattern.valid_postal_code_regexp` `text` - `NOT NULL` 3. `country_postal_code_pattern.clean_postal_code_regexp` `text` 4. `country_postal_code_pattern.clean_postal_code_replace` `text` 5. `country_postal_code_pattern.postal_code_example` `text` - `NOT NULL` 6. `country_postal_code_pattern.postal_code_pattern_checked_on` `date` 7. `country_postal_code_pattern.postal_code_pattern_information_source` `text` 8. `country_postal_code_pattern.postal_code_pattern_belongs_to_pg_xenophile` `boolean` Whether or not this pattern was shipped with the `pg_xenophile` extension. Make sure that, for your custom additions to this table, this column is `false`. Even better, though: contribute new or updated postal code patterns upstream, to `pg_xenophile`, so that everybody may profit from your knowledge. Please note, that you will run into problems with dump/restore when you add records to this table from within your own dependent extension set up scripts. - `NOT NULL` - `DEFAULT false` #### Table: `eu_country` The `eu_country` table has 3 attributes: 1. `eu_country.country_code` `country_code_alpha2` - `NOT NULL` - `PRIMARY KEY (country_code)` - `FOREIGN KEY (country_code) REFERENCES country(country_code)` 2. `eu_country.eu_membership_checked_on` `date` 3. `eu_country.eu_country_belongs_to_pg_xenophile` `boolean` - `NOT NULL` - `DEFAULT false` ### Views #### View: `lang_l10n_en` ```sql SELECT lang.lang_code, lang.lang_belongs_to_pg_xenophile, lang_l10n.l10n_lang_code, lang_l10n.l10n_columns_belong_to_extension_name, lang_l10n.l10n_columns_belong_to_extension_version, lang_l10n.name FROM lang LEFT JOIN lang_l10n ON lang.lang_code::text = lang_l10n.lang_code::text AND lang_l10n.l10n_lang_code::text = 'en'::text; ``` #### View: `country_l10n_en` ```sql SELECT country.country_code, country.country_code_alpha3, country.country_code_num, country.calling_code, country.currency_code, country.country_belongs_to_pg_xenophile, country_l10n.l10n_lang_code, country_l10n.l10n_columns_belong_to_extension_name, country_l10n.l10n_columns_belong_to_extension_version, country_l10n.name FROM country LEFT JOIN country_l10n ON country.country_code::text = country_l10n.country_code::text AND country_l10n.l10n_lang_code::text = 'en'::text; ``` ### Routines #### Procedure: `create_l10n_view (name, name, name, lang_code_alpha2, boolean)` Procedure arguments: | Arg. # | Arg. mode | Argument name | Argument type | Default expression | | ------ | ---------- | ----------------------------------------------------------------- | -------------------------------------------------------------------- | ------------------- | | `$1` | `IN` | `table_schema$` | `name` | | | `$2` | `IN` | `base_table$` | `name` | | | `$3` | `IN` | `l10n_table$` | `name` | | | `$4` | `IN` | `lang_code$` | `lang_code_alpha2` | | | `$5` | `IN` | `temp$` | `boolean` | `false` | Procedure-local settings: * `SET search_path TO xeno, public, pg_temp` #### Function: `fkey_guard (regclass, name, anyelement)` Function arguments: | Arg. # | Arg. mode | Argument name | Argument type | Default expression | | ------ | ---------- | ----------------------------------------------------------------- | -------------------------------------------------------------------- | ------------------- | | `$1` | `IN` | `foreign_table$` | `regclass` | | | `$2` | `IN` | `fkey_column$` | `name` | | | `$3` | `IN` | `fkey_value$` | `anyelement` | | Function return type: `anyelement` Function attributes: `STABLE`, `RETURNS NULL ON NULL INPUT`, `PARALLEL SAFE` #### Function: `l10n_table__maintain_l10n_objects()` The `l10n_table__maintain_l10n_objects()` trigger function is meant to actuate changes to the `l10_table` to the actual l10n tables and views tracked by that meta table. Function return type: `trigger` Function-local settings: * `SET search_path TO xeno, public, pg_temp` * `SET pg_xenophile.in_l10n_table_row_trigger TO true` #### Function: `l10n_table__track_alter_table_events()` Function return type: `event_trigger` Function attributes: `SECURITY DEFINER` Function-local settings: * `SET search_path TO xeno, public, pg_temp` * `SET pg_xenophile.in_l10n_table_event_trigger TO true` #### Function: `l10n_table__track_drop_table_events()` Function return type: `event_trigger` Function attributes: `SECURITY DEFINER` Function-local settings: * `SET search_path TO xeno, public, pg_temp` * `SET pg_xenophile.in_l10n_table_event_trigger TO true` #### Function: `l10n_table_with_fresh_ddl (l10n_table)` Function arguments: | Arg. # | Arg. mode | Argument name | Argument type | Default expression | | ------ | ---------- | ----------------------------------------------------------------- | -------------------------------------------------------------------- | ------------------- | | `$1` | `INOUT` | | `l10n_table` | | Function return type: `l10n_table` Function attributes: `STABLE` Function-local settings: * `SET search_path TO xeno, public, pg_temp` #### Function: `pg_xenophile_base_lang_code()` Function return type: `lang_code_alpha2` Function attributes: `STABLE`, `LEAKPROOF` Function-local settings: * `SET pg_readme.include_this_routine_definition TO true` * `SET search_path TO xeno, public, pg_temp` ```sql CREATE OR REPLACE FUNCTION xeno.pg_xenophile_base_lang_code() RETURNS lang_code_alpha2 LANGUAGE sql STABLE LEAKPROOF SET "pg_readme.include_this_routine_definition" TO 'true' SET search_path TO 'xeno', 'public', 'pg_temp' RETURN (COALESCE(current_setting('app.settings.i18n.base_lang_code'::text, true), current_setting('pg_xenophile.base_lang_code'::text, true), 'en'::text))::lang_code_alpha2 ``` #### Function: `pg_xenophile_meta_pgxn()` Returns the JSON meta data that has to go into the `META.json` file needed for PGXN—PostgreSQL Extension Network—packages. The `Makefile` includes a recipe to allow the developer to: `make META.json` to refresh the meta file with the function's current output, including the `default_version`. And indeed, `pg_xenophile` can be found on PGXN: https://pgxn.org/dist/pg_xenophile/ Function return type: `jsonb` Function attributes: `STABLE` #### Function: `pg_xenophile_readme()` Generates a README in Markdown format using the amazing power of the `pg_readme` extension. Temporarily installs `pg_readme` if it is not already installed in the current database. Function return type: `text` Function-local settings: * `SET search_path TO xeno, public, pg_temp` * `SET pg_readme.include_view_definitions TO true` * `SET pg_readme.include_routine_definitions_like TO {test__%}` #### Function: `pg_xenophile_target_lang_codes()` Function return type: `lang_code_alpha2[]` Function attributes: `STABLE`, `LEAKPROOF` Function-local settings: * `SET pg_readme.include_this_routine_definition TO true` * `SET search_path TO xeno, public, pg_temp` ```sql CREATE OR REPLACE FUNCTION xeno.pg_xenophile_target_lang_codes() RETURNS lang_code_alpha2[] LANGUAGE sql STABLE LEAKPROOF SET "pg_readme.include_this_routine_definition" TO 'true' SET search_path TO 'xeno', 'public', 'pg_temp' RETURN (COALESCE(current_setting('app.settings.i18n.target_lang_codes'::text, true), current_setting('pg_xenophile.target_lang_codes'::text, true), '{}'::text))::lang_code_alpha2[] ``` #### Function: `pg_xenophile_user_lang_code()` Function return type: `lang_code_alpha2` Function attributes: `STABLE`, `LEAKPROOF` Function-local settings: * `SET pg_readme.include_this_routine_definition TO true` * `SET search_path TO xeno, public, pg_temp` ```sql CREATE OR REPLACE FUNCTION xeno.pg_xenophile_user_lang_code() RETURNS lang_code_alpha2 LANGUAGE sql STABLE LEAKPROOF SET "pg_readme.include_this_routine_definition" TO 'true' SET search_path TO 'xeno', 'public', 'pg_temp' RETURN (COALESCE(current_setting('app.settings.i18n.user_lang_code'::text, true), current_setting('pg_xenophile.user_lang_code'::text, true), regexp_replace(current_setting('lc_messages'::text), '^([a-z]{2}).*$'::text, ''::text), 'en'::text))::lang_code_alpha2 ``` #### Function: `set_installed_extension_version_from_name()` Sets the installed extension version string in the column named in the second argument for the extension named in the second argument. See the [`test__set_installed_extension_version_from_name()` test procedure](#procedure-test__set_installed_extension_version_from_name) for a working example of this trigger function. This function was lifted from the `pg_utility_trigger_functions` extension version. 1.4.0, by means of copy-paste to keep the number of inter-extension dependencies to a minimum. Function return type: `trigger` Function-local settings: * `SET search_path TO xeno, public, pg_temp` #### Procedure: `test_dump_restore__l10n_table (text)` This procedure is to be called by the `test_dump_restore.sh` and `test_dump_restore.sql` companion scripts, once before `pg_dump` (with `test_stage$ = 'pre-dump'` argument) and once after `pg_restore` (with the `test_stage$ = 'post-restore'`). Procedure arguments: | Arg. # | Arg. mode | Argument name | Argument type | Default expression | | ------ | ---------- | ----------------------------------------------------------------- | -------------------------------------------------------------------- | ------------------- | | `$1` | `IN` | `test_stage$` | `text` | | Procedure-local settings: * `SET search_path TO xeno, public, pg_temp` * `SET plpgsql.check_asserts TO true` ```sql CREATE OR REPLACE PROCEDURE xeno.test_dump_restore__l10n_table(IN "test_stage$" text) LANGUAGE plpgsql SET search_path TO 'xeno', 'public', 'pg_temp' SET "plpgsql.check_asserts" TO 'true' AS $procedure$ declare _en_expected record; _nl_expected record; _pt_expected record; _es_expected record; _en_actual record; _nl_actual record; _pt_actual record; _es_actual record; begin assert test_stage$ in ('pre-dump', 'post-restore'); if test_stage$ = 'pre-dump' then -- Create the table that will be translated. create table test_uni ( uni_abbr text primary key ); insert into l10n_table (base_table_name, l10n_column_definitions, base_lang_code, target_lang_codes) values ( 'test_uni' ,array['name TEXT NOT NULL'] ,'en'::lang_code_alpha2 ,array['nl', 'fr']::lang_code_alpha2[] ); assert to_regclass('test_uni_l10n') is not null, 'The `_l10n` table should have been created as result of the preceding INSERT into the meta table.'; end if; -- Set up the expected data, now that we for sure have the `test_uni_l10n_*` types, -- regardless of `test_stage$`. _en_expected := row('AX-UNI', 'en', 'Axe University')::test_uni_l10n_en; _nl_expected := row('AX-UNI', 'nl', 'Bijl Universiteit')::test_uni_l10n_nl; if test_stage$ = 'pre-dump' then insert into test_uni_l10n_en (uni_abbr, name) values (_en_expected.uni_abbr, _en_expected.name) returning * into _en_actual ; update test_uni_l10n_nl set name = _nl_expected.name where uni_abbr = _en_expected.uni_abbr returning * into _nl_actual ; elsif test_stage$ = 'post-restore' then select * into _en_actual from test_uni_l10n_en where uni_abbr = _en_expected.uni_abbr; select * into _nl_actual from test_uni_l10n_nl where uni_abbr = _nl_expected.uni_abbr; end if; assert _en_expected = _en_actual; assert _nl_expected = _nl_actual; -- -- Go test a dependent extension (that has its own `l10n_table`) now… -- if test_stage$ = 'pre-dump' then create extension l10n_table_dependent_extension; end if; _pt_expected := row( '👋' ,10 ,false ,'pt' ,null ,null ,'tchau' )::subextension_tbl_l10n_pt; _es_expected := row( '👋' ,10 ,false ,'es' ,null ,null ,'adiós' )::subextension_tbl_l10n_es; if test_stage$ = 'pre-dump' then insert into subextension_tbl_l10n_pt (natural_key, base_tbl_col, localized_text) values (_pt_expected.natural_key, _pt_expected.base_tbl_col, _pt_expected.localized_text) returning * into _pt_actual ; update subextension_tbl_l10n_es set localized_text = _es_expected.localized_text where natural_key = _es_expected.natural_key returning * into _es_actual ; elsif test_stage$ = 'post-restore' then select * into _pt_actual from subextension_tbl_l10n_pt where natural_key = _pt_expected.natural_key; select * into _es_actual from subextension_tbl_l10n_es where natural_key = _es_expected.natural_key; end if; assert _pt_actual = _pt_expected, format('%s != %s', _pt_actual, _pt_expected); assert _es_actual = _es_expected, format('%s != %s', _es_actual, _es_expected); -- -- Test with the dependent subextension and a Portuguese row that has been inserted during installation… -- _pt_expected := row( '👍' ,null ,true ,'pt' ,'l10n_table_dependent_extension' ,'forever' -- Yes, this is a version string. ,'bem' )::subextension_tbl_l10n_pt; _es_expected := row( '👍' ,null ,true ,'es' ,null ,null ,'buen' )::subextension_tbl_l10n_es; if test_stage$ = 'pre-dump' then select * into _pt_actual from subextension_tbl_l10n_pt where natural_key = _pt_expected.natural_key; update subextension_tbl_l10n_es set localized_text = _es_expected.localized_text where natural_key = _es_expected.natural_key returning * into _es_actual ; elsif test_stage$ = 'post-restore' then select * into _pt_actual from subextension_tbl_l10n_pt where natural_key = _pt_expected.natural_key; select * into _es_actual from subextension_tbl_l10n_es where natural_key = _es_expected.natural_key; end if; assert _pt_actual = _pt_expected, format('%s != %s', _pt_actual, _pt_expected); assert _es_actual = _es_expected, format('%s != %s', _es_actual, _es_expected); end; $procedure$ ``` #### Procedure: `test__l10n_table()` Procedure-local settings: * `SET search_path TO xeno, public, pg_temp` * `SET client_min_messages TO WARNING` * `SET plpgsql.check_asserts TO true` * `SET pg_readme.include_this_routine_definition TO true` ```sql CREATE OR REPLACE PROCEDURE xeno.test__l10n_table() LANGUAGE plpgsql SET search_path TO 'xeno', 'public', 'pg_temp' SET client_min_messages TO 'WARNING' SET "plpgsql.check_asserts" TO 'true' SET "pg_readme.include_this_routine_definition" TO 'true' AS $procedure$ declare _row record; _l10n_table l10n_table; begin -- Create the table that will be translated. create table test_uni ( id bigint primary key generated always as identity ,uni_abbr text not null unique -- We need to have more than one non-PK column, to ensure that we're hitting the requirement to agg. -- Also, let's put a space in the column name, so that we're testing proper quoting as well. ,"student rating" bigint default 5 ); <> begin -- This tests that the trigger(s) on `l10n_table` tries to create the `_l10n_nl`-suffixed view -- only once and doesn't crash because of trying to create it twice. insert into l10n_table (base_table_name, l10n_column_definitions, base_lang_code, target_lang_codes) values ( 'test_uni' ,array['name TEXT NOT NULL', '"description (short)" TEXT NOT NULL'] ,'nl'::lang_code_alpha2 -- Apologies for the Dutch East India Company mentality. ,array['nl']::lang_code_alpha2[] ); raise transaction_rollback; exception when transaction_rollback then end with_redundant_target_lang; -- Register `test_uni` with the meta table, to activate all the l10n magic. insert into l10n_table (base_table_name, l10n_column_definitions, base_lang_code, target_lang_codes) values ( 'test_uni' ,array['name TEXT NOT NULL', '"description (short)" TEXT NOT NULL'] ,'nl'::lang_code_alpha2 -- Apologies for the Dutch East India Company mentality. ,array['en', 'fr']::lang_code_alpha2[] ); assert to_regclass('test_uni_l10n') is not null, 'The `_l10n` table should have been created as result of the preceding INSERT into the meta table.'; assert ( select array_agg(pg_class.relname order by pg_class.relname)::name[] from pg_catalog.pg_class where pg_class.relkind = 'v' and pg_class.relnamespace = current_schema::regnamespace and pg_class.relname like 'test\_uni\_l10n\___' ) = array['test_uni_l10n_en', 'test_uni_l10n_fr', 'test_uni_l10n_nl']::name[] ,'3 `_l10n_`-suffixed views should have been created, one for the base language' || ' and 2 for the target languages.'; <> declare _nl_expected record := row( 1, 'AX-UNI', 5, 'nl', 'Bijl Universiteit', 'De trainingsleider in bijlonderhoud en gebruik' )::test_uni_l10n_nl; _en_expected record := row( 1, 'AX-UNI', 5, 'en', 'Axe University', 'The leader in axe maintenance and usage training' )::test_uni_l10n_en; begin insert into test_uni_l10n_nl (uni_abbr, name, "description (short)") values (_nl_expected.uni_abbr, _nl_expected.name, _nl_expected."description (short)") returning * into _row ; assert _row = _nl_expected, format( 'The `RETURNING` clause did not return the data as inserted; %s ≠ %s' ,_row, _nl_expected ); assert _nl_expected = (select row(tbl.*)::test_uni_l10n_nl from test_uni_l10n_nl as tbl), 'The `RETURNING` clause should have returned the same row data as this separate `SELECT`.'; assert exists(select from test_uni_l10n_en where id = _nl_expected.id), 'Even though the row for English doesn''t exist in `test_uni_l10n` yet, it should exist in the' ' `test_l10n_en` view, with NULL values for all the l10n columns.'; update test_uni_l10n_en set "name" = _en_expected."name" ,"description (short)" = _en_expected."description (short)" where id = _nl_expected.id returning * into _row ; assert found, 'The `UPDATE` should have found a row to update in the `test_uni_l10n_en` view.'; assert _row = _en_expected, format('%s ≠ %s; the `RETURNING` clause did not return the data as upserted.', _row, _en_expected); assert _en_expected = (select row(tbl.*)::test_uni_l10n_en from test_uni_l10n_en as tbl), 'The `RETURNING` clause should have returned the same row data as this separate `SELECT`.'; end upsert_into_l10n_lang_view; <> declare _expected_id bigint := currval('test_uni_id_seq') + 1; _nl_expected record := row( _expected_id, 'SIMP-UNI', 2, 'nl', 'Simpschool', 'Simpen voor Elon en Jeff' )::test_uni_l10n_nl; begin insert into test_uni_l10n_nl (id, uni_abbr, "student rating", name, "description (short)") values ( _nl_expected.id ,_nl_expected.uni_abbr ,_nl_expected."student rating" ,_nl_expected.name ,_nl_expected."description (short)" ) returning * into _row ; raise assert_failure using message = 'It should not be possible to explicitly specify a PK value on insert' 'if that PK is `GENERATED ALWAYS`.'; exception when generated_always then end try_to_override_generated_pk; <> declare _expected_id bigint := currval('test_uni_id_seq') + 1; _nl_expected record := row( _expected_id, 'PO-UNI', 7, 'nl', 'Poep-Universiteit', 'De Beste Plek om Te Leren Legen' )::test_uni_l10n_nl; _en_expected record := row( _expected_id, 'PO-UNI', 7, 'en', 'Pooversity', 'The Best Place To Empty Yourself' )::test_uni_l10n_nl; begin insert into test_uni_l10n_nl (uni_abbr, "student rating", name, "description (short)") values ( _nl_expected.uni_abbr ,_nl_expected."student rating" ,_nl_expected.name ,_nl_expected."description (short)" ) returning * into _row ; -- Test that the trigger `test_uni_l10n_en` does an INSERT instead of an UPDATE if no row for this -- PK + lang_code combo exists yet in `test_uni_l10n`. update test_uni_l10n_en set uni_abbr = _en_expected.uni_abbr ,name = _en_expected.name ,"description (short)" = _en_expected."description (short)" where id = _en_expected.id returning * into _row ; assert _row = _en_expected, format('%s ≠ %s', _row, _en_expected); end insert_instead_of_update_on_missing_l10n_record; <> declare _expected_id bigint := currval('test_uni_id_seq') + 1; _fr_expected record := row( _expected_id, 'MOI-UNI', null, 'fr', 'Moiversitee', 'La Premier Bla' )::test_uni_l10n_fr; begin insert into test_uni_l10n_fr (uni_abbr, name, "description (short)") values (_fr_expected.uni_abbr, _fr_expected.name, _fr_expected."description (short)") returning * into _row ; delete from test_uni_l10n_fr where id = _row.id; assert found; assert not exists (select from test_uni where id = _row.id), 'The base table record should have been deleted.'; assert not exists ( select from test_uni_l10n where id = _row.id and l10n_lang_code = _row.l10n_lang_code ) ,'The l10n record should have been deleted, via the `ON DELETE CASCADE` to the base table.'; end delete_via_l10n_view; <> begin alter table test_uni_l10n add description2 text; update test_uni_l10n set description2 = 'Something to satisfy NOT NULL'; -- Because we want to make it NOT NULL. alter table test_uni_l10n alter column description2 set not null; select * into _l10n_table from l10n_table where base_table_name = 'test_uni'; assert _l10n_table.l10n_column_definitions[3] = 'description2 text NOT NULL', 'The `l10n_table__track_alter_table_events` event trigger should have updated the list of l10n' ' columns.'; assert exists( select from pg_attribute where attrelid = 'test_uni_l10n_fr'::regclass and attname = 'description2' ), 'The `description2` column should have been added to the view.'; alter table test_uni_l10n drop column description2 cascade; select * into _l10n_table from l10n_table where base_table_name = 'test_uni'; assert array_length(_l10n_table.l10n_column_definitions, 1) = 2, 'The dropped column should have been removed from the `l10n_table` meta table.'; assert not exists( select from pg_attribute where attrelid = 'test_uni_l10n_nl'::regclass and attname = 'description2' ), 'The `description2` column should have disappeared from the views.'; alter table test_uni add non_l10n_col int not null default 6; select * into _l10n_table from l10n_table where base_table_name = 'test_uni'; assert _l10n_table.base_column_definitions[4] = 'non_l10n_col integer NOT NULL DEFAULT 6', 'The `l10n_table__track_alter_table_events` event trigger should have updated the list of base' ' columns.'; assert (select non_l10n_col from test_uni_l10n_nl where id = 2) = 6; alter table test_uni drop column non_l10n_col cascade; assert not exists( select from pg_attribute where attrelid = 'test_uni_l10n_nl'::regclass and attname = 'non_l10n_col' ), 'The `non_l10n_col` column should have disappeared from the views.'; <> declare _nl_expected record; begin alter table test_uni add column base_col_with_default text not null default 'I am default.'; alter table test_uni_l10n add column localized_image text not null default 'fallback.png'; select * into _l10n_table from l10n_table where base_table_name = 'test_uni'; assert _l10n_table.base_column_definitions[4] = 'base_col_with_default text NOT NULL DEFAULT ''I am default.''::text', format( 'The `l10n_table__track_alter_table_events` event trigger should have updated the list of' ' base columns; base_column_definitions = ''%s''' ,_l10n_table.base_column_definitions ); assert _l10n_table.l10n_column_definitions[3] = 'localized_image text NOT NULL DEFAULT ''fallback.png''::text', format( 'The `l10n_table__track_alter_table_events` event trigger should have updated the list of' ' l10n columns; l10n_column_definitions = ''%s''' ,_l10n_table.l10n_column_definitions ); -- Now, let's test how the defaults behave on insert… _nl_expected := row( 1, 'HOF', 5, 'I am default.', 'nl', 'Wim Hof', 'De Ijsman', 'fallback.png' )::test_uni_l10n_nl; insert into test_uni_l10n_nl (uni_abbr, name, "description (short)") values (_nl_expected.uni_abbr, _nl_expected.name, _nl_expected."description (short)") returning * into _row ; assert _row.base_col_with_default = _nl_expected.base_col_with_default, 'Default should have propegated from the base table to view.'; assert _row.localized_image = _nl_expected.localized_image, 'Default should have propegated from the l10n table to view.'; end add_base_column_with_default_value; <> begin drop table test_uni cascade; assert not exists (select from l10n_table where base_table_name = 'test_uni'); raise transaction_rollback; -- I could have used any error code, but this one seemed to fit best. exception when transaction_rollback then end drop_base_table; end trigger_alter_table_events; -- DELETE-ing the meta info for our l10n table should cascade cleanly, without crashing. delete from l10n_table where base_table_regclass = 'test_uni'::regclass; assert to_regclass('test_uni_l10n') is null, 'The actual `_l10n` table should have been removed when deleting the meta row from `l10n_table`.'; <> declare _expected record; begin -- Let's make a table with a natural primary key that is _not_ `GENERATED ALWAYS`. create table test_species ( scientific_name text primary key -- Just so you know: without the `year_first_described` column, the `INSERT INTO l10n_table` -- would not trigger a certain bug, so please do not allow regressions to occur by removing -- this column. ,year_first_described int ); -- Register `test_species` with the meta table, to activate all the l10n magic. insert into l10n_table (base_table_name, l10n_column_definitions, base_lang_code) values ('test_species' ,'{common_name TEXT}' ,'en') ; insert into test_species_l10n_en (scientific_name, common_name, year_first_described) values ('Taraxacum officinale', 'common dandelion', 1753) ; <> begin insert into test_species_l10n_en (scientific_name, common_name) values ('Taraxacum officinale', 'uncommon dandelion') ; raise assert_failure using message = 'Duplicating a primary key shouldn''t have been possible.'; exception when unique_violation then end insert_duplicate_natural_key; end insert_natural_key; raise transaction_rollback; -- I could have used any error code, but this one seemed to fit best. exception when transaction_rollback then end; $procedure$ ``` #### Function: `updatable_l10n_view()` Function return type: `trigger` Function-local settings: * `SET search_path TO xeno, public, pg_temp` ### Types The following extra types have been defined _besides_ the implicit composite types of the [tables](#tables) and [views](#views) in this extension. #### Domain: `currency_code` Using this domain instead of its underlying `text` type ensures that only uppercase, 3-letter currency codes are allowed. It does _not_ enforce that the `currency_code` exists in the `currency` table. ```sql CREATE DOMAIN currency_code AS text CHECK ((VALUE ~ '^[A-Z]{3}$'::text)); ``` #### Domain: `country_code_alpha2` Using this domain instead of its underlying `text` type ensures that only 2-letter, uppercase country codes are allowed. ```sql CREATE DOMAIN country_code_alpha2 AS text CHECK ((VALUE ~ '^[A-Z]{2}$'::text)); ``` #### Domain: `lang_code_alpha2` ISO 639-1 two-letter (lowercase) language code. ```sql CREATE DOMAIN lang_code_alpha2 AS text CHECK ((VALUE ~ '^[a-z]{2}$'::text)); ``` #### Domain: `lang_code_alhpa3` ISO 639-2/T, ISO 639-2/B, or ISO 639-3 (lowercase) language code. ```sql CREATE DOMAIN lang_code_alhpa3 AS text CHECK ((VALUE ~ '^[a-z]{3}$'::text)); ``` ## Missing/planned/possible features * Currently (as of version 0.7.4), only ISO 639-1 (2-letter) language codes are supported. It would be nice if at least ISO 639-2 3-letter code would be supported, and possibly ISO 639-2/T and 639-2/B as well. Even better would be if [BPC 47 / RFC 5646](https://datatracker.ietf.org/doc/html/rfc5646) was supported. If I (Rowan) do change the primary language identification method, I will try to do so _before_ `pg_xenophile` 1.0 is released, because introducing breaking changes post-1.0 is assholish towards the couple of users that might by then already depend on this extension. ## Extension authors and contributors * [Rowan](https://www.bigsmoke.us/) originated this extension in 2022 while developing the PostgreSQL backend for the [FlashMQ SaaS MQTT cloud broker](https://www.flashmq.com/). Rowan does not like to see himself as a tech person or a tech writer, but, much to his chagrin, [he _is_](https://blog.bigsmoke.us/category/technology). Some of his chagrin about his disdain for the IT industry he poured into a book: [_Why Programming Still Sucks_](https://www.whyprogrammingstillsucks.com/). Much more than a “tech bro”, he identifies as a garden gnome, fairy and ork rolled into one, and his passion is really to [regreen and reenchant his environment](https://sapienshabitat.com/). One of his proudest achievements is to be the third generation ecological gardener to grow the wild garden around his beautiful [family holiday home in the forest of Norg, Drenthe, the Netherlands](https://www.schuilplaats-norg.nl/) (available for rent!). ## Colophon This `README.md` for the `pg_xenophile` extension was automatically generated using the [`pg_readme`](https://github.com/bigsmoke/pg_readme) PostgreSQL extension.