chore: add SQLFluff to pre-commit hooks

.pre-commit-config.yaml

@@ -36,16 +36,14 @@ repos:
         language: rust
         pass_filenames: false
 
-  #
-  # TODO: This needs a more thorough configuration, e.g. spacing, alignment, capitalization, etc
-  #
-  # https://docs.sqlfluff.com/en/stable/production/pre_commit.html
-  # - repo: https://github.com/sqlfluff/sqlfluff
-  #   rev: 3.3.1
-  #   hooks:
-  #     - id: sqlfluff-fix
-  #       name: sqlfluff-fix
-  #       args: [ "--dialect", "postgres" ]
-  #       files: '^.*\.sql$'
+  - repo: https://github.com/sqlfluff/sqlfluff
+    rev: 3.3.1
+    hooks:
+      - id: sqlfluff-fix
+        # AL07: Avoid table aliases in from clauses and join conditions <- we sometimes do this for long names
+        # LT05: Line is too long <- not useful. mostly triggers for comments
+        # LT12: Files must end with a single trailing newline <- already done by trailing-whitespace
+        args: [--dialect=postgres, "--exclude-rules=AL07,LT05,LT12"]
+        files: \.sql$
 
 # TODO: Add NPM hooks for formatting and linting

demo/db/initdb/01.sql

@@ -3,28 +3,30 @@ create extension postgis;
 
 create table trips
 (
-  vendorid              numeric,
-  pickup_datetime       timestamp,
-  dropoff_datetime      timestamp,
-  passenger_count       numeric,
-  trip_distance         numeric,
-  ratecodeid            numeric,
-  store_and_fwd_flag    text,
-  pulocationid          numeric,
-  dolocationid          numeric,
-  payment_type          numeric,
-  fare_amount           numeric,
-  extra                 numeric,
-  mta_tax               numeric,
-  tip_amount            numeric,
-  tolls_amount          numeric,
-  improvement_surcharge numeric,
-  total_amount          numeric,
-  congestion_surcharge  numeric,
-  airport_fee           numeric
+    vendorid numeric,
+    pickup_datetime timestamp,
+    dropoff_datetime timestamp,
+    passenger_count numeric,
+    trip_distance numeric,
+    ratecodeid numeric,
+    store_and_fwd_flag text,
+    pulocationid numeric,
+    dolocationid numeric,
+    payment_type numeric,
+    fare_amount numeric,
+    extra numeric,
+    mta_tax numeric,
+    tip_amount numeric,
+    tolls_amount numeric,
+    improvement_surcharge numeric,
+    total_amount numeric,
+    congestion_surcharge numeric,
+    airport_fee numeric
 );
 
-create function tilebbox(z integer, x integer, y integer, srid integer DEFAULT 3857) returns geometry
+create function tilebbox(
+    z integer, x integer, y integer, srid integer default 3857
+) returns geometry
 immutable
 language plpgsql
 as $$
@@ -46,14 +48,15 @@ begin
     return ST_Transform(bbox, srid);
   end if;
 end;
-$$
-;
+$$;
 
-create or replace function get_trips(z integer, x integer, y integer, query_params json) returns bytea
-    stable
-    strict
-    parallel safe
-    language plpgsql
+create or replace function get_trips(
+    z integer, x integer, y integer, query_params json
+) returns bytea
+stable
+strict
+parallel safe
+language plpgsql
 as $$
 DECLARE
   bounds GEOMETRY(POLYGON, 3857) := TileBBox(z, x, y, 3857);
@@ -92,7 +95,6 @@ BEGIN
   RETURN res;
 
 END;
-$$
-;
+$$;
 
 alter function get_trips(integer, integer, integer, json) owner to postgres;

demo/db/initdb/03.sql

@@ -1,10 +1,12 @@
 drop table if exists trips_by_hour;
 create table trips_by_hour as
 select
-	pulocationid,
-	count(*) as trips_count,
-	round(avg(total_amount)) trips_price,
-	round(avg(extract(epoch from (dropoff_datetime - pickup_datetime)) / 60))::INTEGER trips_duration,
-	date_trunc('hour', pickup_datetime) as pickup_datetime
+    pulocationid,
+    round(
+        avg(extract(epoch from (dropoff_datetime - pickup_datetime)) / 60)
+    )::INTEGER as trips_duration,
+    count(*) as trips_count,
+    round(avg(total_amount)) as trips_price,
+    date_trunc('hour', pickup_datetime) as pickup_datetime
 from trips
 group by pulocationid, date_trunc('hour', pickup_datetime);

martin/src/pg/scripts/query_available_function.sql

@@ -15,60 +15,94 @@
 --   input_names: a JSON array of input parameter names
 --   input_types: a JSON array of input parameter types
 WITH
-    --
-    inputs AS (
-        -- list of input parameters for each function, returned as a jsonb array [{name: type}, ...]
-        SELECT specific_name,
-               jsonb_agg(COALESCE(parameter_name::text, '_') ORDER BY ordinal_position) as input_names,
-               jsonb_agg(data_type::text ORDER BY ordinal_position)                     as input_types
-        FROM information_schema.parameters
-        WHERE parameter_mode = 'IN'
-          AND specific_schema NOT IN ('pg_catalog', 'information_schema')
-        GROUP BY specific_name),
-    --
-    outputs AS (
-        -- list of output parameters for each function, returned as a jsonb array [{name: type}, ...]
-        SELECT specific_name,
-               jsonb_agg(data_type::text ORDER BY ordinal_position)      as out_params,
-               jsonb_agg(parameter_name::text ORDER BY ordinal_position) as out_names
-        FROM information_schema.parameters
-        WHERE parameter_mode = 'OUT'
-          AND specific_schema NOT IN ('pg_catalog', 'information_schema')
-        GROUP BY specific_name),
-    --
-    comments AS (
-        -- list of all comments associated with the function
-        SELECT pg_namespace.nspname                    AS schema,
-               pg_proc.proname                         AS name,
-               obj_description(pg_proc.oid, 'pg_proc') AS description
-        FROM pg_proc
-                 JOIN pg_namespace ON pg_proc.pronamespace = pg_namespace.oid)
-SELECT routines.specific_schema AS schema,
-       routines.routine_name    AS name,
-       routines.data_type       AS output_type,
-       outputs.out_params       AS output_record_types,
-       out_names                AS output_record_names,
-       inputs.input_types       AS input_types,
-       inputs.input_names       AS input_names,
-       comments.description     AS description
+--
+inputs AS (
+    -- list of input parameters for each function, returned as a jsonb array [{name: type}, ...]
+    SELECT
+        specific_name,
+        jsonb_agg(
+            coalesce(parameter_name::text, '_')
+            ORDER BY ordinal_position
+        ) AS input_names,
+        jsonb_agg(
+            data_type::text
+            ORDER BY ordinal_position
+        ) AS input_types
+    FROM information_schema.parameters
+    WHERE
+        parameter_mode = 'IN'
+        AND specific_schema NOT IN ('pg_catalog', 'information_schema')
+    GROUP BY specific_name
+),
+
+--
+outputs AS (
+    -- list of output parameters for each function, returned as a jsonb array [{name: type}, ...]
+    SELECT
+        specific_name,
+        jsonb_agg(
+            data_type::text
+            ORDER BY ordinal_position
+        ) AS out_params,
+        jsonb_agg(
+            parameter_name::text
+            ORDER BY ordinal_position
+        ) AS out_names
+    FROM information_schema.parameters
+    WHERE
+        parameter_mode = 'OUT'
+        AND specific_schema NOT IN ('pg_catalog', 'information_schema')
+    GROUP BY specific_name
+),
+
+--
+comments AS (
+    -- list of all comments associated with the function
+    SELECT
+        pg_namespace.nspname AS schema, -- noqa: RF04
+        pg_proc.proname AS name, -- noqa: RF04
+        obj_description(pg_proc.oid, 'pg_proc') AS description
+    FROM pg_proc
+    INNER JOIN pg_namespace ON pg_proc.pronamespace = pg_namespace.oid
+)
+
+SELECT
+    routines.specific_schema AS schema, -- noqa: RF04
+    routines.routine_name AS name, -- noqa: RF04
+    routines.data_type AS output_type,
+    outputs.out_params AS output_record_types,
+    outputs.out_names AS output_record_names,
+    inputs.input_types,
+    inputs.input_names,
+    comments.description
 FROM information_schema.routines
-         JOIN inputs ON routines.specific_name = inputs.specific_name
-         LEFT JOIN outputs ON routines.specific_name = outputs.specific_name
-         LEFT JOIN comments ON comments.schema = routines.specific_schema AND comments.name = routines.routine_name
-WHERE jsonb_array_length(input_names) IN (3, 4) -- 3 or 4 input parameters
-  AND lower(input_names ->> 0) IN ('z', 'zoom') -- the first int param is either z or zoom
-  AND input_types ->> 0 = 'integer'
-  AND lower(input_names ->> 1) = 'x'            -- the second int param is x
-  AND input_types ->> 1 = 'integer'
-  AND lower(input_names ->> 2) = 'y'            -- the third param is y
-  AND input_types ->> 2 = 'integer'
-  -- the 4th optional parameter can be any name, and must be either json or jsonb
-  AND (input_types ->> 3 = 'json' OR input_types ->> 3 = 'jsonb' OR (input_types ->> 3) IS NULL)
-  -- the output must be either a single bytea value or a table, with the table row being either [bytea] or [bytea, text]
-  AND (
-        (data_type = 'bytea' AND out_params IS NULL)
-        OR (data_type = 'bytea' AND out_params = '["bytea"]'::jsonb)
-        OR (data_type = 'record' AND out_params = '["bytea"]'::jsonb)
-        OR (data_type = 'record' AND out_params = '["bytea", "text"]'::jsonb)
+INNER JOIN inputs ON routines.specific_name = inputs.specific_name
+LEFT JOIN outputs ON routines.specific_name = outputs.specific_name
+LEFT JOIN
+    comments
+    ON
+        routines.specific_schema = comments.schema
+        AND routines.routine_name = comments.name
+WHERE
+    jsonb_array_length(inputs.input_names) IN (3, 4) -- 3 or 4 input parameters
+    -- the first int param is either z or zoom
+    AND lower(inputs.input_names ->> 0) IN ('z', 'zoom')
+    AND inputs.input_types ->> 0 = 'integer'
+    AND lower(inputs.input_names ->> 1) = 'x'            -- the second int param is x
+    AND inputs.input_types ->> 1 = 'integer'
+    AND lower(inputs.input_names ->> 2) = 'y'            -- the third param is y
+    AND inputs.input_types ->> 2 = 'integer'
+    -- the 4th optional parameter can be any name, and must be either json or jsonb
+    AND (
+        inputs.input_types ->> 3 = 'json'
+        OR inputs.input_types ->> 3 = 'jsonb'
+        OR (inputs.input_types ->> 3) IS NULL
+    )
+    -- the output must be either a single bytea value or a table, with the table row being either [bytea] or [bytea, text]
+    AND (
+        (routines.data_type = 'bytea' AND outputs.out_params IS NULL)
+        OR (routines.data_type = 'bytea' AND outputs.out_params = '["bytea"]'::jsonb)
+        OR (routines.data_type = 'record' AND outputs.out_params = '["bytea"]'::jsonb)
+        OR (routines.data_type = 'record' AND outputs.out_params = '["bytea", "text"]'::jsonb)
     )
 ORDER BY routines.specific_schema, routines.routine_name;