Re: BUG #7899: allow key word as alias in subquery but Can't reference it in outer query
От | Jov |
---|---|
Тема | Re: BUG #7899: allow key word as alias in subquery but Can't reference it in outer query |
Дата | |
Msg-id | CADyrUxPsp8+X4Lv-5wuicczNNtCb++fKopD0vTSNpDWShinh1w@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: BUG #7899: allow key word as alias in subquery but Can't reference it in outer query (Jov <amutu@amutu.com>) |
Список | pgsql-bugs |
can we make a cross reference from doc 4.4.1( http://www.postgresql.org/docs/devel/static/sql-syntax-lexical.html#SQL-SYN= TAX-IDENTIFIERS ) to doc 7.3.2 ( http://www.postgresql.org/docs/devel/static/queries-select-lists.html#QUERI= ES-COLUMN-LABELS ) *to mention that =93AS=94 can make a key word to be a identifier?* I have read the doc 4.4.1 and can't find this info from there so I thin it is a bug. 2013/2/22 Jov <amutu@amutu.com> > I have get the description of this behaviour from doc 7.3.2 > > http://www.postgresql.org/docs/devel/static/queries-select-lists.html#QUE= RIES-COLUMN-LABELS > . > > from the error message now,I think syntax error may emmit by the > parser,parser error make it hard to get the really condition such as the > second token =91end=92 is a select item from a select statement. > > thanks > > 2013/2/22 Tom Lane <tgl@sss.pgh.pa.us> > >> "=3D?ISO-8859-1?B?Sm92?=3D" <amutu@amutu.com> writes: >> > xxx=3D# select end from (select 1 as end) t; >> > ERROR: syntax error at or near "end" >> > LINE 1: select end from (select 1 as end) t; >> > ^ >> >> You need to double-quote the outer use of "end", viz >> >> regression=3D# select "end" from (select 1 as end) t; >> end >> ----- >> 1 >> (1 row) >> >> In the context with "as", Postgres can tell the word is meant as a >> column identifier not a keyword, but there's no way for it to know that >> in the outer usage. >> >> > I think use key word as alias should get a error message such as >> $key_word >> > is Key Word,should not be used as identifier bla bla,so for some compl= ex >> > query error user can save a lot of time. >> >> Well, yours is the first complaint we've ever had in that direction, >> whereas we used to get a lot of complaints in the opposite direction, >> back when we didn't allow keywords to be used for column names. >> So I doubt we'll change it. It is an interesting gotcha though :-( >> >> regards, tom lane >> . >> >> > > > -- > Jov > blog: http:amutu.com/blog <http://amutu.com/blog> > --=20 Jov blog: http:amutu.com/blog <http://amutu.com/blog>
В списке pgsql-bugs по дате отправления: