Copy table with separate sequences for serial columns

While creating a column management tool I came across the need to quickly replicate tables in PostgreSQL so I would not test new tools with non-test tables. To effectively test the new column tool that I ultimately intend to use on the table parts I created this new tool to duplicate parts so I'd end up with a parts1 table. When I thought I had finally gotten all the issues ironed out I encountered the following error when the column tool deletes the table:


ERROR: cannot drop table parts because other objects depend on it
DETAIL: default for table parts1 column id depends on sequence
parts_id_seq1

I've spent the greater part of my day working towards this solution so in short can I simply use string functions to rename the SEQUENCE_NAME variable to disassociate the parts table from the parts table or is it a more convoluted issue than that? Here is the query:

DO $
  DECLARE
    SEQUENCE_NAME VARCHAR;
  BEGIN
    SELECT s.relname INTO SEQUENCE_NAME
    FROM pg_class AS s JOIN pg_depend d ON d.objid = s.oid 
    INNER JOIN pg_class AS t ON d.objid = s.oid AND d.refobjid = t.oid 
    INNER JOIN pg_attribute AS a ON (d.refobjid, d.refobjsubid) = (a.attrelid, a.attnum) 
    INNER JOIN pg_namespace AS n ON n.oid = s.relnamespace 
    WHERE s.relkind = 'S' 
    AND n.nspname = 'public' 
    AND t.relname='parts';

    LOCK TABLE parts;
    CREATE TABLE parts1 (LIKE parts INCLUDING ALL);
    INSERT INTO parts1 SELECT * FROM parts;
    PERFORM setval(SEQUENCE_NAME::regclass, (SELECT max(id) FROM parts)+1);
  END;
$ LANGUAGE plpgsql;

To create a copy as close as possible use INCLUDING ALL with CREATE TABLE .. (LIKE ..) since there can be any number of columns with defaults that you obviously want to copy.

You just want serial columns to get their own, independent sequence, which makes a lot of sense and probably should have been the default behavior to begin with.

Postgres 10 "fixed" that by adding IDENTITY columns complying to the SQL-standard, which have internal, exclusively dedicated sequences, and behave as desired with CREATE TABLE .. (LIKE ..). The manual:

INCLUDING IDENTITY

Any identity specifications of copied column definitions will be
copied. A new sequence is created for each identity column of the
new table, separate from the sequences associated with the old
table.

Bold emphasis mine. Existing serial columns remain unchanged. Consider replacing serial columns. See:

• Auto increment table column

• How to change a table ID from serial to identity?

Function to copy tables with (or without) serial columns

While any serial columns are still involved, this function should do the job:

Copies any given table (must exist) with new given name and independent serial columns (if any).

Data is not included, it's trivial to copy that, too.

CREATE OR REPLACE FUNCTION f_copy_tbl(_tbl regclass, _newtbl text, _newschema text = NULL)
  RETURNS void
  LANGUAGE plpgsql VOLATILE AS
$func$
DECLARE
   _newtbl_qual_n_quot text;
   _sql text;
BEGIN
   -- If target schema is not given, default to schema of source
   IF _newschema IS NULL THEN      
      SELECT INTO _newschema  relnamespace::regnamespace::text
      FROM   pg_class
      WHERE  oid = _tbl;
   END IF;
   
   _newtbl_qual_n_quot := format('%I.%I', _newschema, _newtbl);
   
   -- Copy table
   EXECUTE format('CREATE TABLE %s (LIKE %s INCLUDING ALL);', _newtbl_qual_n_quot, _tbl);

   -- Fix serial columns, if any
   SELECT INTO _sql
          string_agg('CREATE SEQUENCE ' || seq_qual_n_quot, E';\n') || E';\n'
       || string_agg(format('ALTER SEQUENCE %s OWNED BY %s.%I'
                           , seq_qual_n_quot, _newtbl_qual_n_quot, a.attname), E';\n') || E';\n'
       || 'ALTER TABLE ' || _newtbl_qual_n_quot || E'\n  '
       || string_agg(format($ALTER %I SET DEFAULT nextval('%s'::regclass)$
                          , a.attname, seq_qual_n_quot), E'\n, ')
   FROM   pg_attribute  a
   JOIN   pg_attrdef    ad ON ad.adrelid = a.attrelid
                          AND ad.adnum   = a.attnum
        , format('%I.%I'
               , _newschema
               , _newtbl || '_' || a.attname || '_seq'
                ) AS seq_qual_n_quot  -- new seq name, qualified & quoted where necessary
   WHERE  a.attrelid = _tbl
   AND    a.attnum > 0
   AND    NOT a.attisdropped
   AND    a.atttypid = ANY ('{int,int8,int2}'::regtype[])
   AND    pg_get_expr(ad.adbin, ad.adrelid) = 'nextval('''
            || (pg_get_serial_sequence (a.attrelid::regclass::text, a.attname))::regclass
            || '''::regclass)'
   ;

   IF _sql IS NOT NULL THEN
      EXECUTE _sql;
   END IF;
END
$func$;

Call:

SELECT f_copy_tbl('tbl', 'tbl1');

Or:

SELECT f_copy_tbl('myschema.tbl', 'CpY_tbl_odd_name', 'other_schema');

Produces and executes SQL code of the form:

CREATE TABLE tbl1 (LIKE tbl INCLUDING ALL);
-- only if there are serial columns:
CREATE SEQUENCE tbl1_tbl_id_seq;     -- one line per serial type ...
CREATE SEQUENCE "tbl1_Odd_COL_seq";  -- .. two in this example
ALTER SEQUENCE tbl1_tbl_id_seq OWNED BY tbl1.tbl_id;
ALTER SEQUENCE "tbl1_Odd_COL_seq" OWNED BY tbl1."Odd_COL";
ALTER TABLE tbl1
  ALTER tbl_id SET DEFAULT nextval('tbl1_tbl_id_seq'::regclass)
, ALTER "Odd_COL" SET DEFAULT nextval('"tbl1_Odd_COL_seq"'::regclass);

-
The source (1st parameter) must be a table, view, materialized view, composite type, or foreign table. Optionally schema-qualified.

-
The 2nd parameter is the new table name.

-
The 3nd parameter is schema of the new table. If not given, it defaults to the schema of the source.

-
The system column pg_attrdef.adsrc has been dropped in Postgres 12. Using pg_get_expr(ad.adbin, ad.adrelid) instead as instructed in the manual Works in older versions, too.

-
Only serial columns get their own sequence. Other column default are copied unchanged - including nextval() from a sequence that is not owned by the column or differs in any way from a serial.

-
The function is safe against SQL injection and should work with arbitrary table and column names.

db<>fiddle here

Old sqlfiddle

StackExchange Database Administrators Q#91632, answer score: 7