column name corruption

What could cause column name corruption such as this: transid 1555 studentid 18082 courseid 253 institutionid 29 courseofferingid 618 purchasetime 1997-09-24 12:06:02 type P paymentmethod acctnum registrationprice 0.00 salestax B9Ùntate Underneart salestax, it should be, statetax, instead it is that jibberish. Each time we run the query, that column name changes to something else. ?????? Thanks, Floyd

The syscolumns where fine, the oncheck -cc showed no corruption. The data in the table was fine, but could be selected only with a select all, and could not be selected with that columname in the query. It would say that the column was not found. My suspicsion at that point was that something in memory was corrupted, since there was no apparent corruption to syscolumns. So I had them bounce the server, and now the problem appears to be gone. Thanks, Floyd "June C. Hunt" <june_c_hunt@hotmail.com> wrote: "Floyd Welle...." wrote: > >What could cause column name corruption such as this: >transid 1555 >studentid 18082 >courseid 253 >institutionid 29 >courseofferingid 618 >purchasetime 1997-09-24 12:06:02 >type P >paymentmethod >acctnum >registrationprice 0.00 >salestax >B9Ùntate > >Underneart salestax, it should be, statetax, instead it is that jibberish. >Each time we run the query, that column name changes to something else. >?????? You are running this in DB-Access, right? Running the query in menu-mode or non-menu mode? Is it always the twelfth column name that appears off? What about the twelfth column from some other table? Does the table have exactly twelve columns? If it has more, is the thirteenth showing up OK? Does the column name look normal in syscolumns? All questions and no answers, I know, but... that does seem odd. About the only time I can recall seeing something even remotely like this is when I'm editing a file in 'vi' and start scrolling the last line on the screen when there are other lines to display. A simple redraw of the screen clears up the display. I wish I could be of more help, but maybe one of the questions will help lead you a little closer to an answer. -- June Hunt

run the query and pipe the output through od -cx. My knee jerk reaction is that there is garbage in salestax - or statetax which is causing the cursor placement to go awry. j. ----- Original Message ----- From: "Floyd Welle...." <fwellers@yahoo.com> To: <ids@iiug.org> Sent: Thursday, March 17, 2005 2:33 PM Subject: column name corruption [4543] > What could cause column name corruption such as this: > transid 1555 > studentid 18082 > courseid 253 > institutionid 29 > courseofferingid 618 > purchasetime 1997-09-24 12:06:02 > type P > paymentmethod > acctnum > registrationprice 0.00 > salestax > B9Ùntate > > Underneart salestax, it should be, statetax, instead it is that jibberish. Each time we run the query, that column name changes to something else. ?????? > > Thanks, > Floyd > > >

Hi, Try running oncheck -pp table rowid after determining the rowid using the select statement. This will really show the exact data representing the salestax in the page on disk; the data will be in hex. If the data is correct, then it is a tool program or an engine problem when it goes to get the data. How do you access the data (dbaccess, esql/c program, other ...)? If the data is correct, then try to run different selects going through an index and not going through an index (sequentially). You could also try to use SQLIDEBUG (using sqliprint to read it) to see all the data coming back from the engine. You could also try to create a new table with the same structure (indexes as well) and copy the data using insert into new_table select * from current_table. How long is the row? Does it go over the page size? I am trying to ask several questions to find a path to an answer. Regards, Khaled Bentebal ConsultiX Tél: 33 (0) 1 39 72 17 00 Fax: 33 (0) 1 39 72 17 01 Mobile: 33 (0) 6 07 78 41 97 Email: khaled.bentebal@consult-ix.fr Site Web: http://www.consult-ix.fr ----- Original Message ----- From: "Floyd Welle...." <fwellers@yahoo.com> To: <ids@iiug.org> Sent: Thursday, March 17, 2005 8:47 PM Subject: RE: column name corruption [4544] > The syscolumns where fine, the oncheck -cc showed no corruption. > The data in the table was fine, but could be selected only with a select all, and could not be selected with that columname in the query. It would say that the column was not found. > > My suspicsion at that point was that something in memory was corrupted, since there was no apparent corruption to syscolumns. So I had them bounce the server, and now the problem appears to be gone. > > Thanks, > Floyd > > "June C. Hunt" <june_c_hunt@hotmail.com> wrote: > "Floyd Welle...." wrote: > > > >What could cause column name corruption such as this: > >transid 1555 > >studentid 18082 > >courseid 253 > >institutionid 29 > >courseofferingid 618 > >purchasetime 1997-09-24 12:06:02 > >type P > >paymentmethod > >acctnum > >registrationprice 0.00 > >salestax > >B9Ùntate > > > >Underneart salestax, it should be, statetax, instead it is that jibberish. > >Each time we run the query, that column name changes to something else. > >?????? > > You are running this in DB-Access, right? Running the query in menu-mode or > non-menu mode? Is it always the twelfth column name that appears off? What > about the twelfth column from some other table? Does the table have exactly > twelve columns? If it has more, is the thirteenth showing up OK? Does the > column name look normal in syscolumns? > > All questions and no answers, I know, but... that does seem odd. About the > only time I can recall seeing something even remotely like this is when I'm > editing a file in 'vi' and start scrolling the last line on the screen when > there are other lines to display. A simple redraw of the screen clears up > the display. I wish I could be of more help, but maybe one of the questions > will help lead you a little closer to an answer. > > -- > June Hunt > > > > >