Обсуждение: pgaccess - international characters
Hello, I have a Linux server running Debian 2.2 with PostgreSQL 6.5.3. I would really like to avoid upgrading the Postgres installation as this is running on a production machine. But if it solves the problem I will of course do so. Now we have several Windows workstations using pgaccess to enter data into several tables. Everything is fine when using forms. International characters (like the german umlaute) work fine. But when I open the table directly don't get human readable output for the characters. It varies from one or several blank characters to "garbage". A pg_dump on the Linux server shows the same results. But well this is not so bad, as the stuff can be done using the forms I have written. Now we want to get the data into an Excel 97 spreadsheet. This shows the same behaviour as opening the tables directly. And I don't know what to do about it. To test it I wrote some records into the table using MS access. This seems to work. The characters are shown correctly in the pgaccess table view and the Excel sheet. Even a pg_dump produces human readable characters. I have created the database with several different encodings: LATIN1, WIN, ALT and SQL_ASCII. I also tried several combinations with the encodings and executing a SET CLIENT_ENCODING = "..." in pgaccess. But it didn't make a difference. I'm using the ODBC driver v06-50-0000 as my primitive excel sql-function did not work with version 07_01_0006. Although it still showed the same symptons with SQL queries using MS query. How can I convince pgaccess to write the characters in a readable format into the database? We were using an older pgaccess and tcl version, but I upgraded to the latest pgaccess and the latest point release of tcl8.3. Thanks for any help, Lars -- Those of you who think you know everything are annoying those of us who do.
> Now we have several Windows workstations using pgaccess to enter data into > several tables. Everything is fine when using forms. International > characters (like the german umlaute) work fine. But when I open the table > directly don't get human readable output for the characters. It varies > from one or several blank characters to "garbage". You can try TkSql. I wrote it for an Italian company, we use all sort of accented letters from latin1, it uses TkTable instead of the canvas, maybe that makes the difference (forms use entries that in turn use different fonts) > Now we want to get the data into an Excel 97 spreadsheet. This shows the > same behaviour as opening the tables directly. And I don't know what to do > about it. You'll also find a function that allows you to "print on Excel" what you have on the table (writes out a file and 'start' it or calls Excel on it, needs two variable to configure). You'll also find filter and search functions, and if you know a little bit of tcl/tk you can use direcly the main tcl "bricks" that have planty of options, hooks for validity checks, inhibit the view of parts of the tables and so on. You can find more information on TkSql at http://pcco2.mi.infn.it/sd/soft A rish demo is included in the demo. Simple install script + .rpm, and .deb license is GPL. Hope this helps, if you have any problems w/ TkSql, please let me know, I didn't have much feed-back since I released it (2 months) ago. I have being using TkSql since a couple of years. sandro dentella *:-) PS. it you are concerned about rewriting some forms, it is as difficoult as writing: set attrs { first_name last_name address - address2 - phone fax } editDb -attrs $attrs-tbl names -mask -- Sandro Dentella *:-) e-mail: sandro@ermit.it sandro.dentella@mi.infn.it http://pcco2.mi.infn.it/sd/soft
Hello, On Fri, Jul 13, 2001 at 01:32:22AM +0200, Lars Bensmann wrote: > > Now we have several Windows workstations using pgaccess to enter data into > several tables. Everything is fine when using forms. International > characters (like the german umlaute) work fine. But when I open the table > directly don't get human readable output for the characters. It varies > from one or several blank characters to "garbage". I tested this on another box with Debian unstable installed (this includes PostgreSQL 7.1.2). When I use psql to enter umlaut characters everything is fine. pg_dump and gtksql show proper results. Even pgaccess displays the characters correctly. But when I enter the characters in pgaccess or update a record I entered with psql I get this "garbage" again. I examined this a bit more closely and it seems that the characters are translated like this: "a: <C3><A4> "o: <C3><B6> "u: <C3><BC> "A: <C3><84> "O: <C3><96> "U: <C3><9C> "s: <C3><9F> I get the same results when I enter the data using Tksql as suggested by Sandro. This seems to be related to tcl or tk. Am I the only one with this problem? What can I do about this? TIA, Lars -- Any time things appear to be going better, you have overlooked something.