Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> export/import puzzler
Platform: Ora 8.0.5 EE and 8.1.7 EE on NT
Given an 8.0.5 database, we took a full export with the following parms:
userid=SYSTEM/******@WRPTDB
FULL=Y
FILE=F:\ORAEXP\WRPT\WRPTFULL.DMP
LOG=F:\ORAEXP\WRPT\FULL.LOG
RECORDLENGTH=4096
BUFFER=20480
ROWS=Y
GRANTS=Y
INDEXES=Y
COMPRESS=N
CONSTRAINTS=Y
Then we installed 8.1.7 on a clean system, built the db (but did not create a
user schema) and ran an import with the following parm file:
userid=system/******@wrptdb
FROMUSER=NMM
TOUSER=NMM
FILE=F:\ORAEXP\WRPT\WRPTFULL.DMP
LOG=F:\ORAEXP\WRPT\ImpNMM.LOG
RECORDLENGTH=4096
BUFFER=40960
ROWS=Y
GRANTS=Y
INDEXES=Y
IGNORE=Y
COMMIT=Y
On the first attempt, the first 2 tables loaded fine, but the third started
throwing
IMP-00019: row rejected due to ORACLE error 1400 IMP-00003: ORACLE error 1400 encountered ORA-01400: cannot insert NULL into ("NMM"."CAL_DAY"."CAL_DAY_NME")
Very puzzling, seeing as how the table definition that created the NOT NULL constraint came from the same export that created the supposedly null data.
Knowing it wasn't a real solution, we truncated the 3 tables that had been touched, removed the NOT NULL constraint from the problem table, and re-ran the import. This time it got to the 3d table, pegged the CPU, and seemed to just sit there. I left it overnight, then killed it this morning. (It didn't respond to ctrl-C, so I used NT Task Mgr to knock it down.) This action did not allow the process to close the log file, so it still shows a length of zero and is thus unreadable. A SELECT COUNT on the problem table returned a count of over 38 MILLION rows, when we know the original only had around 14,000 rows.
So . . . . lets
SELECT
COMMENTS,
IDEAS,
CRITICISM,
SOLUTION
FROM
NEWSGROUP_PARTICIPANTS;
-- Ed Stevens (Opinions expressed do not necessarily represent those of my employer.)Received on Tue May 14 2002 - 08:15:16 CDT
![]() |
![]() |