Re: serial not accepted as datatype in ALTER TABLE ... ALTER COLUMN - Mailing list pgsql-hackers

From Andy Fan
Subject Re: serial not accepted as datatype in ALTER TABLE ... ALTER COLUMN
Date
Msg-id 87o7cegopq.fsf@163.com
Whole thread Raw
In response to serial not accepted as datatype in ALTER TABLE ... ALTER COLUMN  (Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>)
Responses Re: serial not accepted as datatype in ALTER TABLE ... ALTER COLUMN
List pgsql-hackers
Hi Ashutosh, 

> data_type is described on that page as "Data type of the new column,
> or new data type for an existing column." but CREATE TABLE
> documentation [2] redirects data_type to [3], which mentions serial.
> The impression created by the documentation is the second statement
> above is a valid statement as should not throw an error; instead
> change the data type of the column (and create required sequence).

I didn't find out a reason to not support it, if have any reason, I
think it is better have some explaination in the document.

> In code ATPrepAlterColumnType() calls typenameTypeIdAndMod(), whereas
> transformColumnDefinition() (called for ALTER TABLE ... ADD COLUMN and
> CREATE TABLE) handles "serial" data type separately. Looks like we are
> missing a call to transformColumnDefinition() in
> transformAlterTableStmt() under case AT_AlterColumnType.

I tried your idea with the attatchment, it is still in a drafted state
but it can be used as a prove-of-concept and for better following
communicating.  Just one point needs to metion is serial implies
"default value" + "not null" constaint. So when we modify a column into
serial, we need to modify the 'NULL value' and only to the default value
at the RewriteTable stage.

-- 
Best Regards
Andy Fan


Attachment

pgsql-hackers by date:

Previous
From: Li Japin
Date:
Subject: Re: Thoughts about NUM_BUFFER_PARTITIONS
Next
From: Sutou Kouhei
Date:
Subject: Re: Why is pq_begintypsend so slow?