Re: [SQL] combining columns in select
От | Herouth Maoz |
---|---|
Тема | Re: [SQL] combining columns in select |
Дата | |
Msg-id | l03130302b4265990a282@[147.233.159.109] обсуждение исходный текст |
Ответ на | Re: [SQL] combining columns in select (Doug Thistlethwaite <doug.t@deltanet.com>) |
Список | pgsql-sql |
At 07:04 +0200 on 10/10/1999, Doug Thistlethwaite wrote: > What I was trying to do is develop a java application using the ODBC and then > move it to postgres & a linux database when I had everything working. > Unfortunately, I have discovered that there are many differences in the > "standard SQL" that each driver uses. > > I could not find any ODBC drivers that us local data that allow the "||" > concatenation symbol. Most of them use the "+" symbol. I have found the >same > problem with "Memo" vs "text" fields as well. The ODBC driver for dbase (the > one which seems to work best in my case) also uses the # symbol for dates > while postgres likes the ' symbol. > > I guess I have to live with this and write separate query statements for each > implementation. If you want to write a cross-platform JDBC application, you must avoid non-standard SQL. The SQL92 standard operator for string concatenation is just a single pipe char |. So neither "+" nor "||" are standard. In this case, I don't think either of the databases supports the actual standard. You can probably overcome this difficulty by setting a variable to either "||" or "+", according to the type of driver you get, and build the SQL queries using this variable. As for the other things, you really have to conform to the standard. Neither Memo nor Text are standard types. If you want standard, stick to CHARACTER VARYING. As for dates, there are escape sequences in JDBC which are supposed to convert dates from a common format to the format acceptable to your particular database. Use these, and you are "standard". Herouth -- Herouth Maoz, Internet developer. Open University of Israel - Telem project http://telem.openu.ac.il/~herutma
В списке pgsql-sql по дате отправления: