Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.misc -> Corrupt data causes application to take over CPU
I have been working with a table that has a corrupted record in
it. I describe it as corrupted because the data is so bizarre. For
example, fields defined as NUMBER(12) have values like -1.297E-73
and -6.90E+125 in them. A couple of date fields have values like
'00-DECEMB' in them.
So, the record's pretty screwed up. But the thing that I find so bizarre is that when I select one particular field from this record, the CPU goes to 99% and the statement does not return. So if I issue the following very simple command from sqlplus:
SELECT CONFERENCEID FROM CHANNELINCONFERENCE WHERE ID = 889911; sqlplus takes over the machine in the way I just described. BTW, CONFERENCEID is the field with the data in it. If I issue the command:
SELECT CONFERENCEID FROM CHANNELINCONFERENCE WHERE ID != 889911; everything is fine.
I've never experienced this before. I would not have thought that extracting one piece of data would cause this. Does anyone have a clue as to what is wrong? I can delete the record and handle the problem that way, I'm just curious as to the possible causes of this.
BTW, I can dump the field. The command:
SELECT DUMP(CONFERENCEID, 16) FROM CHANNELINCONFERENCE WHERE ID = 889911; yields:
Typ=2 Len=9: 3,1f,1,7,78,67,4,9,3
although the same command for the other records yield results like:
Typ=2 Len=4: c3,5f,1a,32 Typ=2 Len=4: c3,5f,1a,4c Typ=2 Len=4: c3,5f,1e,28 Typ=2 Len=4: c3,5f,21,34
I would have expected the dump of the field for the corrupted record to be similar, i.e. len of 4, not 9.
I appreciate any suggestions as to what might cause this.
Jim Lyons
DBA
University of Texas at Austin
Received on Tue Feb 11 2003 - 10:33:12 CST