RETURNING

With a serial or IDENTITY column, you can return the generated ID from the same command:

INSERT INTO tbl(filename)
VALUES ('my_filename')
RETURNING tbl_id;

See:

• Auto increment table column
• Basics in the manual.

Explicitly fetch value

If filename needs to include tbl_id (redundantly), you can still use a single query.

Use lastval() or the more specific currval():

INSERT INTO tbl (filename)
VALUES ('my_filename' || currval('tbl_tbl_id_seq')   -- or lastval()
RETURNING tbl_id;

See:

• Reference value of serial column in another column during same INSERT

If multiple sequences may be advanced in the process (even by way of triggers or other side effects) the sure way is to use currval('tbl_tbl_id_seq').

pg_get_serial_sequence() gets the sequence name

If you don't know the sequence name:

INSERT INTO tbl (filename)
VALUES ('my_filename' || currval(pg_get_serial_sequence('tbl', 'tbl_id'))
RETURNING tbl_id;

"tbl_tbl_id_seq" is the default name for a table tbl with a serial column tbl_id. But there are no guarantees. A column default can fetch values from any sequence if so defined. Also, if the default name is taken when creating the table, Postgres picks the next free name according to a simple algorithm.

tbl_tbl_id_seq must be visible in the the current search_path, else you must schema-qualify.

fiddle

The previously obtained value of a sequence is accessed with the currval() function.

But that will only return a value if nextval() has been called before that.

There is absolutely no way of "peeking" at the next value of a sequence without actually obtaining it.

But your question is unclear. If you call nextval() before doing the insert, you can use that value in the insert. Or even better, use currval() in your insert statement:

select nextval('my_sequence') ...

... do some stuff with the obtained value

insert into my_table(id, filename)
values (currval('my_sequence'), 'some_valid_filename');

Answer for 2022

You can use pg_sequence_last_value() if you know the name of the sequence:

SELECT pg_sequence_last_value('public.person_id_seq');

I stumbled upon this question b/c I was trying to find the next sequence value by table. This didn't answer my question however this is how its done, and it may help those looking for the sequence value not by name but by table:

SELECT nextval(pg_get_serial_sequence('<your_table>', 'id')) AS new_id; 

Hope it helps :)

If your are not in a session you can just nextval('you_sequence_name') and it's just fine.

To answer your question literally, here's how to get the next value of a sequence without incrementing it:

SELECT
 CASE WHEN is_called THEN
   last_value + 1
 ELSE
   last_value
 END
FROM sequence_name

Obviously, it is not a good idea to use this code in practice. There is no guarantee that the next row will really have this ID. However, for debugging purposes it might be interesting to know the value of a sequence without incrementing it, and this is how you can do it.

I tried this and it works perfectly

@Entity
public class Shipwreck {
  @Id
  @GeneratedValue(strategy = GenerationType.SEQUENCE, generator = "seq")
  @Basic(optional = false)
  @SequenceGenerator(name = "seq", sequenceName = "shipwreck_seq", allocationSize = 1)
  Long id;

....

CREATE SEQUENCE public.shipwreck_seq
    INCREMENT 1
    START 110
    MINVALUE 1
    MAXVALUE 9223372036854775807
    CACHE 1;