Postgresql

CREATE TABLE if not exists "o"."A"
(
    --编号
    "Number" bigint NOT NULL,
     --……
     --创建时间
    "CreateTime" timestamp without time zone NOT NULL DEFAULT now(),
    --更新时间
    "UpdateTime" timestamp without time zone NOT NULL DEFAULT now()
)
partition by range("CreateTime");

CREATE UNIQUE INDEX IF NOT EXISTS u_idx_AdsARH_CT_N
ON "o"."A"("CreateTime","Number");
CREATE INDEX if not exists idx_AdsARH_AdsCId_AdsAId 
ON "orders"."A" USING btree ("CId","AId");


--add partition
DO 
$do$
DECLARE
	r_year record;
	r_month record;
    v_year integer;
    v_month integer;
    v_startData date;
	v_endData date;
    str_sql text;
    str_next_month text;
    v_schema text:='o';
    v_table_preffix_name text:='A';
	v_years int[]:= '{2015,2016,2017,2018,2019,
    2020,2021,2022,2023,2024,2025,2026,2027,2028,2029,
    2030,2031,2032,2033,2034,2035,2036,2037,2038,2039}';
    v_months int[]:='{1,2,3,4,5,6,7,8,9,10,11,12}';
BEGIN

   for r_year in (
		WITH data AS (
			SELECT v_years::text[] AS arr
		)
		SELECT arr[i] AS "v_year"
			FROM data, generate_subscripts((SELECT arr FROM data), 1) i
	)
	LOOP
		RAISE NOTICE 'start.... ';
		v_year=CAST(coalesce(r_year."v_year", '0') AS integer);	
		RAISE NOTICE 'v_year: % ', v_year;
        continue when v_year is null;

		for r_month in (
            WITH data AS (
                SELECT v_months::int[] AS arr
            )
            SELECT arr[i] AS "v_month"
                FROM data, generate_subscripts((SELECT arr FROM data), 1) i
        )
        LOOP
            v_month= CAST(coalesce(r_month."v_month", '0') AS integer);
            RAISE NOTICE 'v_month: % ', v_month;
            continue when v_month is null;

            v_startData=to_date(concat(v_year,v_month), 'yyyymm');
            RAISE NOTICE 'v_startData: %,%,%', 
			v_startData,to_char(v_startData,'YYYYMMDD'),to_char(v_startData,'YYYY-MM-DD');
            v_endData=v_startData + interval '1 month';
            RAISE NOTICE 'v_endData: %,%', v_endData,to_char(v_endData,'YYYY-MM-DD');
            continue when v_startData is null;
            continue when v_endData is null;
            
			str_sql = format('create table if not exists  "%I".%I 
            partition of "%I".%I 
            for values from (''%I'') to(''%I'')', 
            v_schema, concat(v_table_preffix_name,'_',to_char(v_startData,'YYYYMMDD')),
			v_schema,v_table_preffix_name,
            v_startData,v_endData);
            RAISE NOTICE 'str_sql: % ', str_sql;
			
			execute str_sql;

        END LOOP;
		str_sql = format('create table if not exists  "%I".%I 
            partition of "%I".%I 
            DEFAULT', 
            v_schema, concat(v_table_preffix_name,'_','default'),
			v_schema,v_table_preffix_name);
        RAISE NOTICE 'str_sql: % ', str_sql;
        execute str_sql;
		RAISE NOTICE 'end.... ';
	END LOOP;
END;
$do$;	

　　

建立分区表的时候（Postgresql 不建议建立主键，这样会引起以后必要的麻烦，如果需要可以建立唯一索引）

主键约束或唯一约束必须包含分区字段。这样才能确保整个分区表内的唯一性，因为每个分区上的唯一约束只维护自身的唯一性。