Re: How to copy rows into same table efficiently

On 10/26/21 1:04 AM, Arun Suresh wrote:
>
> Dear PG experts,
>
> We have a tenant discriminator column in our tables to enable storage of
> data from multiple
> tenants. This column is also part of the composite primary key.
> The customers may request creation of a tenant copy, which means if they
> currently have a
> tenant id "ABCD", they would like to copy all data "where tenant_id =
> 'ABCD'" to a new copy tenant "XYZ".
> The copy must also be done on the same table.
>
> Current approach taken is to build a query like below:
> INSERT INTO mytable (col1, col2, col3, col4) SELECT col1, 'XYZ', col3,
> col4 FROM mytable WHERE col2 = 'ABCD'
>
> Is there a better way to do this?
> There could be other tables with foreign key reference, would a simple
> ordering of the copy based on table relationship suffice?

That would be my first thought.  Of course, FK dependencies quickly become a
rat's nest, so beware.

> Also if the table has millions of records, what are the things to take
> care of?

Take care of?

Anyway, for millions of rows, I might use COPY instead of INSERT (depending
on how many millions, how many indices, how large the rows, how fast the
machine, etc.

--
Angular momentum makes the world go 'round.