Sqlldr date format

Question

Dear experts,
I'm trying to load txt file into table via sqlldr. I'm receiving error "Rejected - Error on table TEST_TABLE, column PERIOD_START_TIME. ORA-01843: not a valid month
The date format in flat txt file is MM:DD:YYYY; in the output table I want to have is DD.MM.YYYY
It looks that sqlldr wrongly interprets input date format - when the input is 07-08-2012 then the output is 07.08.2012 so it assumes that the input is in format MM-DD-YYYY instead of DD-MM-YYYY. And when the day is higher than 12 (13,14 etc) then an error appears
Do you know where/how to force SQLLDR to interpret it correcly? Is it somehow related to NLS_Lang. settings?
Control file looks as follow:
OPTIONS (SKIP=1)
LOAD DATA
INFILE test.txt
APPEND
INTO TABLE TEST_TABLE
FIELDS TERMINATED BY ';'
TRAILING NULLCOLS
(NE_ID,
LAC_ID,
PERIOD_START_TIME DATE 'DD.MM.YYYY HH24:MI:SS',
PERIOD_DURATION,
NSCURRENT,
NSAVERAGE
Table definition:
NE_ID     NUMBER     
LAC_ID     NUMBER     
PERIOD_START_TIME     DATE     
PERIOD_DURATION     NUMBER     
NSCURRENT     NUMBER
NSAVERAGE     NUMBER
INPUT DATA:
NE_ID;LAC_ID;PERIOD_START_TIME;PERIOD_DURATION;NSCURRENT;NSAVERAGE;
576527001;37000;07-16-2012 09:00:00;60;24846;24956;
576527001;37000;07-08-2012 10:00:00;60;1;1
Thanks in advance for any hints
Rgds
Lukasz

Answer 1

Hi, Lukasz,
Welcome to the forum!
user8640349 wrote:
Dear experts,
I'm trying to load txt file into table via sqlldr. I'm receiving error "Rejected - Error on table TEST_TABLE, column PERIOD_START_TIME. ORA-01843: not a valid month
The date format in flat txt file is MM:DD:YYYY; in the output table I want to have is DD.MM.YYYYIn the table, if the column is a DATE, then it won't have either format. Format, such as 'MM:DD:YYYY' versus 'DD.MM.YYYY' only applies to strings.
It looks that sqlldr wrongly interprets input date format - when the input is 07-08-2012 then the output is 07.08.2012 so it assumes that the input is in format MM-DD-YYYY instead of DD-MM-YYYY. And when the day is higher than 12 (13,14 etc) then an error appears
Do you know where/how to force SQLLDR to interpret it correcly? Is it somehow related to NLS_Lang. settings?
Control file looks as follow:
OPTIONS (SKIP=1)
LOAD DATA
INFILE test.txt
APPEND
INTO TABLE TEST_TABLE
FIELDS TERMINATED BY ';'
TRAILING NULLCOLS
(NE_ID,
LAC_ID,
PERIOD_START_TIME DATE 'DD.MM.YYYY HH24:MI:SS',The format you give after the keyword DATE tessls what the input data looks like. If the month comes before the day in your input file, then you should be saying
{code}
PERIOD_START_TIME DATE 'MM.DD.YYYY HH24:MI:SS',
{code}
PERIOD_DURATION,
NSCURRENT,
NSAVERAGE
Table definition:
NE_ID     NUMBER     
LAC_ID     NUMBER     
PERIOD_START_TIME     DATE     
PERIOD_DURATION     NUMBER     
NSCURRENT     NUMBER
NSAVERAGE     NUMBER
INPUT DATA:
NE_ID;LAC_ID;PERIOD_START_TIME;PERIOD_DURATION;NSCURRENT;NSAVERAGE;
576527001;37000;07-16-2012 09:00:00;60;24846;24956;
576527001;37000;07-08-2012 10:00:00;60;1;1SQL*Loader only controls how data gets into a table. It has nothing to do with what you do with that data after it is in the table.
If you want a DATE column displayed in a particular format, then use TO_CHAR in the query that produces the display, or change NLS_DATE_FORMAT to set the default date format in your session. For example:
{code}
ALTER SESSION SET NLS_DATE_FORMAT = 'DD.MM.YYYY HH24:MI:SS';
SELECT ename
,     hiredate
FROM     scott.emp
ORDER BY hiredate
{code}
Output:
{code}
ENAME HIREDATE
SMITH 17.12.1980 00:00:00
ALLEN 20.02.1981 00:00:00
WARD 22.02.1981 00:00:00
JONES 02.04.1981 00:00:00
BLAKE 01.05.1981 00:00:00
CLARK 09.06.1981 00:00:00
TURNER 08.09.1981 00:00:00
MARTIN 28.09.1981 00:00:00
KING 17.11.1981 00:00:00
JAMES 03.12.1981 00:00:00
FORD 03.12.1981 00:00:00
MILLER 23.01.1982 00:00:00
SCOTT 19.04.1987 00:00:00
ADAMS 23.05.1987 00:00:00
{code}
It doesn't matter what format the date information was in before it was entered into the table. In the table, all DATEs have the same format.