Обсуждение: timestamp 'default' question

All,

I have a TABLE entry for time:

 timeaccessioned  | timestamp without time zone | not null default
('now'::text)::timestamp(6) with time zone


w/ default set to 'now', I thought postgres would fill in time - not so?

When I attempt to load feature table w/out time entered ->

psql:feature.stdin:1: ERROR:  copy: line 1, Missing data for column
"timeaccessioned"


COPY feature (feature_id, dbxref_id, organism_id, name, uniquename,
residues, seqlen, md5checksum, type_id, timeaccessioned,
timelastmodified) FROM stdin;
1       \N      133     scaffold_1      chlre01.1       AAATCCAAGC
376197  53adae6396a767c7e88910c73b9d0b6b        3
\.


Can I get postgres to load current time, or do I need to generate it in
the script creating the SQL?

Charles

> w/ default set to 'now', I thought postgres would fill in time - not
> so?

Try now()

CREATE TABLE users (
     creation_date timestamp DEFAULT 'now()' NOT NULL
);

Jeff

Charles Hauser <chauser@duke.edu> writes:
> When I attempt to load feature table w/out time entered ->
> psql:feature.stdin:1: ERROR:  copy: line 1, Missing data for column
> "timeaccessioned"

Well, yes, because your COPY command is claiming that all eleven columns
will be supplied in the input:

> COPY feature (feature_id, dbxref_id, organism_id, name, uniquename,
> residues, seqlen, md5checksum, type_id, timeaccessioned,
> timelastmodified) FROM stdin;
> 1       \N      133     scaffold_1      chlre01.1       AAATCCAAGC
> 376197  53adae6396a767c7e88910c73b9d0b6b        3
> \.

I think what you probably intend is

COPY feature (feature_id, dbxref_id, organism_id, name, uniquename,
residues, seqlen, md5checksum, type_id) FROM stdin;

which tells COPY to read those columns from the input and insert
defaults for the remaining columns.

            regards, tom lane