从csv复制到具有ID序列列自动递增PSQL的表

Question

寻找一种解析解决方案，该解决方案将从CSV复制到我的数据库中，并生成最终的序列ID。 要么在后端，要么在复制。 照原样，我的ID在复制后为空。

echo CREATE TABLE IF NOT EXISTS qs_facts_template (id SERIAL NOT NULL, symbol varchar(8), timestamp date, close real null, open real null, high real null, low real null, volume real null, CONSTRAINT qs_facts_template_key PRIMARY KEY (id)) WITH (OIDS=TRUE) TABLESPACE pg_default; ALTER TABLE qs_facts_template OWNER to postgres; | psql -U postgres -h %host% readyloop

echo CREATE TABLE IF NOT EXISTS qs_facts AS select * from qs_facts_template;| psql -U postgres -h %host% readyloop

echo \copy qs_facts (symbol, timestamp, close, open, high, low, volume) from PROGRAM 'cat C:\Users\user\Documents\quantshare\quotes.csv' DELIMITER ';' CSV HEADER| psql -U postgres -h %host% %dbName%

这篇文章： https : //unix.stackexchange.com/questions/277080/copy-csv-data-while-simultanely-filling-serial-column无法解决我的问题。 我使用了Guido提到的命令并使副本起作用，但是id均为null。

我在想也许可以在复制后设置值。

Answer 1

您可以编写一个BEFORE触发器：

CREATE FUNCTION id_trigger() RETURNS trigger
   LANGUAGE plpgsql AS
$$BEGIN
   NEW.id = nextval(pg_get_serial_sequence(TG_ARGV[0], 'id'));
   RETURN NEW;
END;$$;

CREATE TRIGGER id_trigger BEFORE INSERT ON qs_facts_template
   FOR EACH ROW EXECUTE PROCEDURE id_trigger('qs_facts_template');

Answer 2

问题是复制不会复制除列以外的所有内容，这增加了您无法将表更改为串行的事实。

因此，我从原始模板表中将脚本导出到pg4admin中，并能够导出此脚本

    echo DROP materialized view if exists mv_qs_facts cascade| psql -U postgres -h %host% %dbName%
    echo DROP materialized view if exists mv_qs_symbols cascade| psql -U postgres -h %host% %dbName%

    echo DROP TABLE if exists qs_facts cascade| psql -U postgres -h %host% %dbName%

    echo drop table if exists qs_facts_template CASCADE;| psql -U postgres -h %host% readyloop
Rem create qs_fact table    
    echo CREATE TABLE IF NOT EXISTS qs_facts_template (id SERIAL, symbol varchar(8), timestamp date, close real null, open real null, high real null, low real null, volume real null, CONSTRAINT qs_facts_template_key PRIMARY KEY (id)) WITH (OIDS=TRUE) TABLESPACE pg_default; ALTER TABLE qs_facts_template OWNER to postgres; | psql -U postgres -h %host% readyloop

    echo CREATE TABLE IF NOT EXISTS qs_facts AS select * from qs_facts_template; | psql -U postgres -h %host% readyloop

    echo ALTER TABLE public.qs_facts ADD CONSTRAINT qs_facts_key PRIMARY KEY (id); | psql -U postgres -h %host% readyloop

    REM I looked at psqladmin and exported the sequence from qs_facts_template
    REM http://www.postgresqltutorial.com/postgresql-serial/
    echo CREATE SEQUENCE public.qs_facts_id_seq INCREMENT 1 START 1 MINVALUE 1 MAXVALUE 9223372036854775807 CACHE 1; ALTER SEQUENCE public.qs_facts_id_seq OWNER TO postgres;| psql -U postgres -h %host% readyloop

    echo ALTER TABLE public.qs_facts ALTER COLUMN id SET DEFAULT nextval('qs_facts_id_seq'); | psql -U postgres -h %host% readyloop