[英]Looping through tables to retrieve max dates from a created date column in Postgres
I need to monitor the ongoing flow of data using the latest created date for a set of tables.我需要使用一组表的最新创建日期来监视正在进行的数据流。 Basically, I need to batch run
基本上,我需要批量运行
SELECT MAX(z_date_creation)
FROM table_schema.table_name
on a set of tables I retrieve with在我检索的一组表上
SELECT
c.table_schema,
c.table_name
FROM information_schema."columns" c
WHERE c.column_name LIKE '%z_date_creation'
AND c.table_schema = 'datawarehouse'
AND c.table_name NOT LIKE 'partition%'
and then feed it to a dedicated "ods.dates_derniere_maj" table I'll be pluging reports on.然后将其提供给专用的“ods.dates_derniere_maj”表,我将在其中插入报告。
I am using a cursor for like of a better idea to iterate through the tables in need to get the MAX(z_date_creation)
from.我正在使用 cursor 作为一个更好的主意来遍历需要从中获取
MAX(z_date_creation)
的表。 I manage to feed the table_schema
and table_name
values into my ods.dates_derniere_maj
table but cannot find a way to also get the MAX(z_date_creation)
from these tables.我设法将
table_schema
和table_name
值输入到我的ods.dates_derniere_maj
表中,但找不到从这些表中获取MAX(z_date_creation)
的方法。
I'm stuck with the nested query part.我被嵌套查询部分困住了。
Here is what I've come up with so far:到目前为止,这是我想出的:
DO $$
DECLARE
table_rec record ;
max_date TEXT DEFAULT NOW();
cursor1 CURSOR FOR
SELECT DISTINCT c.table_schema, c.table_name, c.column_name
FROM information_schema."columns" c
WHERE c.table_schema = 'datawarehouse'
AND c.table_name NOT LIKE 'partition%'
AND c.column_name LIKE '%creation%';
from_clause TEXT;
date_column TEXT;
BEGIN
FOR table_rec IN cursor1
LOOP
from_clause := CONCAT(table_rec.table_schema, '.', table_rec.table_name);
date_column := CONCAT(table_rec.table_schema, '.', table_rec.table_name,'.','z_date_creation');
Which code herebelow ?
PREPARE nom_req (text, text) AS
SELECT MAX($1) FROM $2 ; ---> not working, syntax error on $2
max_date := EXECUTE nom_req (date_column, from_clause) ; ---> not working
SELECT MAX(date_column) INTO max_date FROM CONCAT(from_clause) ; ----> not working
INSERT INTO ods.dates_derniere_maj (schema_name, table_name, z_date_creation_max)
VALUES (table_rec.table_schema, table_rec.table_name, max_date);
END LOOP;
END $$;`
I have tried passing the table_rec.table_schema and table_rec.table_name variables directly in the FROM clause but that didn't work so I tried to concatenate them beforehand.我曾尝试在 FROM 子句中直接传递 table_rec.table_schema 和 table_rec.table_name 变量,但这没有用,所以我尝试事先连接它们。
Any help will be much appreciated !任何帮助都感激不尽 !
Thanks a bunch !非常感谢!
Franck法兰克
Try something like this:尝试这样的事情:
CREATE FUNCTION max_date() RETURNS date LANGUAGE plpgsql AS
$$
DECLARE
table_rec record ;
max_date date ;
result date ;
cursor1 CURSOR FOR
SELECT DISTINCT c.table_schema, c.table_name, c.column_name
FROM information_schema."columns" c
WHERE c.table_schema = 'datawarehouse'
AND c.table_name NOT LIKE 'partition%'
AND c.column_name LIKE '%creation%';
BEGIN
FOR table_rec IN cursor1
LOOP
EXECUTE FORMAT( 'SELECT max(%I) FROM %I.%I'
, table_rec.column_name
, table_rec.table_schema
, table_rec.table_name
)
INTO max_date ;
result = greatest(result, max_date) ;
END LOOP ;
RETURN result ;
END ;
$$ ;
Here is what I came up with with Edouard's invaluable help !这是我在 Edouard 的宝贵帮助下得出的结论!
DO $$
DECLARE
table_rec record ;
max_date TEXT DEFAULT NOW();
cursor1 CURSOR FOR SELECT DISTINCT c.table_schema, c.table_name, c.column_name
FROM information_schema."columns" c
WHERE c.table_schema = 'datawarehouse'
AND c.table_name NOT LIKE 'partition%'
AND c.column_name LIKE '%creation%';
BEGIN
FOR table_rec IN cursor1
LOOP
EXECUTE FORMAT( 'SELECT max(%I) FROM %I.%I'
, table_rec.column_name
, table_rec.table_schema
, table_rec.table_name
)
INTO max_date ;
INSERT INTO ods.dates_derniere_maj (schema_name, table_name, z_date_creation_max)
VALUES (table_rec.table_schema, table_rec.table_name, max_date);
END LOOP;
END $$;
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.