too many dotted names
Hi,
Does it make sense to treat these ?
ALTER TABLE "s'd"".s'd"""."s's'd""." ADD COLUMN id bigint DEFAULT
nextval('"s''d".s''d""."s''d".d"s''"');
ERROR: improper relation name (too many dotted names): s'd.s'd"".s'd.d"s'"
SQL state: 42601
PostgreSQL 9.1devel on i686-pc-linux-gnu, compiled by GCC gcc
(Ubuntu/Linaro 4.5.1-7ubuntu2) 4.5.1, 32-bit
Best regards,
Vladimir Kokovic, DP senior, Belgrade, Serbia
On Wed, Apr 6, 2011 at 4:23 PM, Vladimir Kokovic
<vladimir.kokovic@gmail.com> wrote:
Hi,
Does it make sense to treat these ?
ALTER TABLE "s'd"".s'd"""."s's'd""." ADD COLUMN id bigint DEFAULT
nextval('"s''d".s''d""."s''d".d"s''"');ERROR: improper relation name (too many dotted names): s'd.s'd"".s'd.d"s'"
SQL state: 42601
Treat them as what?
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
On 4/7/11, Robert Haas <robertmhaas@gmail.com> wrote:
On Wed, Apr 6, 2011 at 4:23 PM, Vladimir Kokovic
<vladimir.kokovic@gmail.com> wrote:Hi,
Does it make sense to treat these ?
ALTER TABLE "s'd"".s'd"""."s's'd""." ADD COLUMN id bigint DEFAULT
nextval('"s''d".s''d""."s''d".d"s''"');ERROR: improper relation name (too many dotted names): s'd.s'd"".s'd.d"s'"
SQL state: 42601Treat them as what?
Even nextval('"s''d".s''d""."s''d".d"s''"') is correct literal,
PostgreSQL parser treats them like error.
I think that only solution is to prohibit ". (double quote and dot) to
be part of schema or sequence name.
Best regards,
Vladimir Kokovic, DP senior, Belgrade, Serbia
Vladimir Kokovic <vladimir.kokovic@gmail.com> writes:
On 4/7/11, Robert Haas <robertmhaas@gmail.com> wrote:
On Wed, Apr 6, 2011 at 4:23 PM, Vladimir Kokovic
<vladimir.kokovic@gmail.com> wrote:ALTER TABLE "s'd"".s'd"""."s's'd""." ADD COLUMN id bigint DEFAULT
nextval('"s''d".s''d""."s''d".d"s''"');
ERROR: improper relation name (too many dotted names): s'd.s'd"".s'd.d"s'"
Treat them as what?
Even nextval('"s''d".s''d""."s''d".d"s''"') is correct literal,
Really? According to whom? This works for me:
regression=# create schema "s'd"".s'd""";
CREATE SCHEMA
regression=# create table "s'd"".s'd"""."s's'd""." (f1 int);
CREATE TABLE
regression=# create sequence "s'd"".s'd"""."s's'd"".s" ;
CREATE SEQUENCE
regression=# ALTER TABLE "s'd"".s'd"""."s's'd""." ADD COLUMN id bigint DEFAULT nextval('"s''d"".s''d"""."s''s''d"".s"');
ALTER TABLE
I think you've made up some theory about how to quote funny characters
in nextval's argument, and it's a wrong theory. You have to double
single quotes because you're writing a string literal, but other than
that it should look just like a quoted identifier in SQL.
regards, tom lane