Transfer data between databases with PostgreSQL

半腔热情 提交于 2019-11-29 19:40:35
Nate

I just had to do this exact thing so I figured I'd post the recipe here. This assumes that both databases are on the same server.

First, copy the table from the old db to the new db (because apparently you can't move data between databases). At the commandline:

pg_dump -U postgres -t <old_table> <old_database> | psql -U postgres -d <new_database>

# Just adding extra space here so scrollbar doesn't hide the command

Next, grant permissions of the copied table to the user of the new database. Log into psql:

psql -U postgres -d <new_database>

ALTER TABLE <old_table> OWNER TO <new_user>;

\q

Finally, copy data from the old table to the new table. Log in as the new user and then:

INSERT INTO <new_table> (field1, field2, field3) 
SELECT field1, field2, field3 from <old_table>;

Done!

leonbloy

Databases are isolated in PostgreSQL; when you connect to a PostgreSQL server you connect to just one database, you can't copy data from one database to another using a SQL query.

If you come from MySQL: what MySQL calls (loosely) "databases" are "schemas" in PostgreSQL - sort of namespaces. A PostgreSQL database can have many schemas, each one with its tables and views, and you can copy from one schema to another with the schema.table syntax.

If you really have two distinct PostgreSQL databases, the common way of transferring data from one to another would be to export your tables (with pg_dump -t ) to a file, and import them into the other database (with psql).

If you really need to get data from a distinct PostgreSQL database, another option - mentioned in Grant Johnson's answer - is dblink, which is an additional module (in contrib/).

This worked for me to copy a table remotely from my localhost to Heroku's postgresql:

pg_dump -C -t source_table -h localhost source_db | psql -h destination_host -U destination_user -p destination_port destination_db

This creates the table for you.

For the other direction (from Heroku to local) pg_dump -C -t source_table -h source_host -U source_user -p source_port source_db | psql -h localhost destination_db

From: hxxp://dbaspot.c om/postgresql/348627-pg_dump-t-give-where-condition.html (NOTE: the link is now broken)

# create temp table with the data
psql mydb
CREATE TABLE temp1 (LIKE mytable);
INSERT INTO temp1 SELECT * FROM mytable WHERE myconditions;
\q

# export the data to a sql file
pg_dump --data-only --column-inserts -t temp1 mtdb > out.sql
psql mydb
DROP TABLE temp1;
\q

# import temp1 rows in another database
cat out.sql | psql -d [other_db]
psql other_db
INSERT INTO mytable (SELECT * FROM temp1);
DROP TABLE temp1;

Another method useful in remotes

  # export a table csv and import in another database
  psql-remote> COPY elements TO '/tmp/elements.csv' DELIMITER ',' CSV HEADER;
  $ scp host.com:/tmp/elements.csv /tmp/elements.csv
  psql-local> COPY elements FROM '/tmp/elements.csv' DELIMITER ',' CSV;

There are three options for copying it if this is a one off:

  1. Use a db_link (I think it is still in contrib)
  2. Have the application do the work.
  3. Export/import

If this is an ongoing need, the answers are:

  1. Change to schemas in the same DB
  2. db_link

Just like leonbloy suggested, using two schemas in a database is the way to go. Suppose a source schema (old DB) and a target schema (new DB), you can try something like this (you should consider column names, types, etc.):

INSERT INTO target.Awards SELECT * FROM source.Nominations;
Anvesh

You can not perform a cross-database query like SQL Server; PostgreSQL does not support this.

The DbLink extension of PostgreSQL is used to connect one database to another database. You have install and configure DbLink to execute a cross-database query.

I have already created a step-by-step script and example for executing cross database query in PostgreSQL. Please visit this post: PostgreSQL [Video]: Cross Database Queries using the DbLink Extension

francisco

Actually, there is some possibility to send a table data from one PostgreSQL database to another. I use the procedural language plperlu (unsafe Perl procedural language) for it.

Description (all was done on a Linux server):

  1. Create plperlu language in your database A

  2. Then PostgreSQL can join some Perl modules through series of the following commands at the end of postgresql.conf for the database A:

    plperl.on_init='use DBI;'
    plperl.on_init='use DBD::Pg;'
    
  3. You build a function in A like this:

    CREATE OR REPLACE FUNCTION send_data( VARCHAR )
    RETURNS character varying AS
    $BODY$
    my $command = $_[0] || die 'No SQL command!';
    my $connection_string =
    "dbi:Pg:dbname=your_dbase;host=192.168.1.2;port=5432;";
    $dbh = DBI->connect($connection_string,'user','pass',
    {AutoCommit=>0,RaiseError=>1,PrintError=>1,pg_enable_utf8=>1,}
    );
    my $sql = $dbh-> prepare( $command );
    eval { $sql-> execute() };
    my $error = $dbh-> state;
    $sql-> finish;
    if ( $error ) { $dbh-> rollback() } else {  $dbh-> commit() }
    $dbh-> disconnect();
    $BODY$
    LANGUAGE plperlu VOLATILE;
    

And then you can call the function inside database A:

SELECT send_data( 'INSERT INTO jm (jm) VALUES (''zzzzzz'')' );

And the value "zzzzzz" will be added into table "jm" in database B.

易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!