SYSDATE+(60*24)/1440

SYSDATE+(60*24)/1440
What is the exact date and time of this conversion?

user8674910 wrote:
hi,
my sysdate is 18-jan-10 01:00:00
based on this formula the output is given below: (SYSDATE+(60*24)/1440 )
20-jan-10 13:00:00 P.M
60 is hr
24 is min ( 24 * 60 sec =1440 seconds )
1440 is minutesYou appear to have done something wrong if you truly got that output. You also appear to have a few misconceptions over what the above formula represents.
Date differences in Oracle are measured in terms of the number of days. In order to convert that into hours, you have to multiply by 24, to convert that into minutes, you have to multiply by 24*60 (1440) and to convert to seconds, multiply by 24*60*60.
eg:
(date1 - date2) = diff_in_days
=> date1 = date2 + diff_in_days
diff_in_hours = diff_in_days * 24
diff_in_minutes = diff_in_days * 24 * 60
diff_in_seconds = diff_in_days * 24 * 60 * 60
=> diff_in_days = diff_in_hours / 24 = diff_in_minutes / (24 * 60) = diff_in_seconds / (24 * 60 * 60)Therefore when you want to add minutes onto a date, you would use the "diff_in_minutes / (24 * 60)" format, and in the OP's example, that works out to be 1440 / 1440 which is the same as saying 1 day.
So your result can't possibly have been correct; if you think it was, please recreate in SQL*Plus and copy the resultant commands, like:
SQL> alter session set nls_date_format='dd/mm/yyyy hh24:mi:ss';
Session altered.
SQL> select sysdate, sysdate + (60*24)/1440 from dual;
SYSDATE SYSDATE+(60*24)/144
18/01/2010 11:27:20 19/01/2010 11:27:20
1 row selected.

Similar Messages

How to find out the rows inserted in the last n minutes ?

Hi all,
I have a log table where, every time a warning is issued, a new row is added. The table has a field with the sysdate when the warning was issued.
I'd need to collect all warnings in the last (supposing) 30 minutes and send a mail to the administrator. Looking at some faqs I have elaborated this:
select * from PROCEDURE_LOGS p where ((p.DATE_LOG - sysdate)*-1440) < 30
However this does not produce anything.
Can anybody help me to correct this statement ?
Thanks a lot
Frank

Hi, Frank,
That looks okay to me. Are you sure you have rows in the table with a date_log in the last 30 minutes? Post some sample data (CREATE TABLE and INSERT statements), and the results you expect from that data for a given value of SYSDATE.
By the way, you can make the WHERE clause clearer like this:
select     *
from      PROCEDURE_LOGS     p
where      ( (sysdate - p.DATE_LOG)
     * 1440           -- = 24 * 60
     ) < 30but this won't change the results.
Edited by: Frank Kulash on Dec 1, 2010 9:51 AM
I just saw Toon's reply:
where     p.DATE_LOG > sysdate - (30 / 1440)as Toon suggested, is clearer still, and potentially more efficient if ther's an index on date_log.

Oarcle Lite cannot subtract negative numbers from date

Hello
we have a very strange problem. We have a query that need to convert times between different timezones. So, we have
In Oracle lite, the query
select sysdate - (-60/1440) nd from dual returns sysdate + 2 - 60/1440.
Whenever a negative fraction is subtracted from date, it always adds two days and then adds the fraction. Looks like negative of a negative fraction is not handled correctly in oracle lite.
Did anyone encounter this problem? We have the system in production and wondering if anyone found a solution.
Thanks

HI sd,
I suspect there's a simpler way, but this works, as seen in the example. Start and finish times are both incremented by one day as the data and formula are filled down the columns to show the effect of crossing the weekend days. Formula (as it appears in row 7) is shown above the table. The Row 2 version is copied below. Cels in column D are formatted as shown in the Inspector.
D2: =NETWORKDAYS(B2,C2,)+DURATION(,,24*(DUR2DAYS(C2-B2)-INT(DUR2DAYS(C2-B2))),,,)
Regards,
Barry

Need To use SQL Code to Appt Dates 2 Days from current Getting Error

I am trying to pull data that shows appt information on a patient:
FName,LName, PhoneNumber, ApptDate, ApptTime, and Status that equals Active. I need to pull this info everyday. How can I set this up in a stored procedure to pull all this information automatically on a daily basis where it pulls by Appt Date that is 2 days greater than the current date?
SELECT
     Patient.First,
     Patient.Last,
     Patient.HomeArea,
     Patient.HomePhone,
     Patient.Status,
     ApptHis.Date,
     ApptHis.Time
From
     Patient, ApptHis
Where
     Status= 'Active' and
     ApptHis.ApptHisDate >= dateadd(day,datediff(day,0,getDate())+2,0)     and
     ApptHis.ApptHisDate < dateadd(day,datediff(day,0,getDate())+3,0)
Please Help I am getting an ORA-00900 Error.
Thanks!

In your select statement you specify the column ApptHis.Time but it looks like in the database the actual column name is Time1.
Also, Date is a reserve word in Oracle. It's a really bad idea to create columns using reserve words. You may have to use double quotes (i.e. "Date") to specify the column name Date in your selects.
Try this:
SELECT
Patient.First,
Patient.Last,
Patient.HomeArea,
Patient.HomePhone,
Patient.Status,
ApptHis.Date,
ApptHis.Time1
From
Patient, ApptHis
Where
Status= 'Active' and
ApptHis.ApptHisDate between trunc(sysdate+2) and trunc(sysdate+3)-1/1440;
NumbNutz

Need to retrieve last 10 minutes records from database

Hi ,
I am Srinivas, I want to retrieve calls done in last 10 min for that i had a query
select count(*) as a from ONLINE_BALANCE where DATETIME >= SYSDATE - INTERVAL '10/1440' HOUR;
It is giving a error;
Can you suggest me with necessary modifications.

Hi ,
I am Srinivas, I want to retrieve calls done in last
10 min for that i had a query
select count(*) as a from ONLINE_BALANCE where
DATETIME >= SYSDATE - INTERVAL '10/1440' HOUR;
It is giving a error;
Can you suggest me with necessary modifications.
... where DATETIME >= SYSDATE - INTERVAL '10' MINUTE;RTFM:
http://download.oracle.com/docs/cd/B19306_01/server.102/b14200/sql_elements003.htm#sthref372
pratz

SQL -Reply

We use Dynamic SQL for returning several rows as well as only one
row. Set up a textdata variable and use it's concat() method to construct
the desired SQL statement:
DynamicSQLStatement: TextData = new ;
DynamicSQLStatement.SetValue('select id_num, ') ;
DynamicSQLStatement.Concat('lst_name, ') ;
DynamicSQLStatement.Concat('fst_name, ') ;
DynamicSQLStatement.Concat('brth_dte, ') ;
DynamicSQLStatement.Concat('ssn, ') ;
---------------------- If SSN entered -----------------------
if SSN.isDigit then
DynamicSQLStatement.Concat('where ssn = ') ;
DynamicSQLStatement.Concat(SSN.IntegerValue);
DynamicSQLStatement.Concat(' ');
else
if NamePattern then
DynamicSQLStatement.Concat('where lst_name like ');
DynamicSQLStatement.Concat('''').Concat(HoldName).Concat('%').Concat(''' ') ;
CheckFirstName() ;
CheckBirthdate() ;
else
DynamicSQLStatement.Concat('where lst_name = ') ;
DynamicSQLStatement.Concat('''').Concat(HoldName).Concat(''' ') ;
CheckFirstName() ;
CheckBirthdate() ;
end if ;
end if ;
and so on.... Read up on DBSession class to prepare sql, open, fetch, &
close cursors.
dynStatement = var_session.Prepare (commandString = SQLselect,
inputDataSet=inputDescriptor,
cmdType =
statementType) ;
begin
rowType = var_session.OpenCursor(statementHandle =
dynStatement,
inputDataSet =
inputDescriptor,
resultDataSet =
outputDescriptor) ;
while true do
numRows = var_session.FetchCursor(statementHandle =
dynStatement,
resultDataSet = outputData) ;
if numRows <= 0 then
exit ;
else
myobj.lst_name=TextData(outputData.GetValue(position = 2)) ;
myobj.fst_name=TextData(outputData.GetValue(position = 3)) ;
myobj.brth_dte=DateTimeData(outputData.GetValue(position = 4)) ;
myobj.ssn=IntegerNullable(outputData.GetValue(position = 5)) ;
etc, etc...
To retrieve just one row...
numRows = var_session.Select(statementHandle = dynStatement,
inputDataSet =inputDescriptor,
resultDataSet= outputData,
maxRows=1) ;
if numRows <= 0 then
ReturnCount.Value = 0;
else
ReturnCount=IntegerData(outputData.GetValue(position=1)
end if ;
Let me know if I'm way off base here. :)
Steven Barnes
IDPA
MOTTA 11/19/96 07:32am >>>Forte Users,
I would like to know if its possible to build an SQL select at run
time. That is, to concat to a textdata and run as SQL that textdata.
If its not possible, is there any other ideia of how we could do
it?

with rt as
(select 1 sno ,sysdate sdate from dual union all
select 2,sysdate - (1/(1440*60)) from dual union all
select 3,sysdate - (1/(1440)) from dual union all
select 4,sysdate - (30/(1440)) from dual union all
select 5,sysdate -1+(1/(1440*60)) from dual union all
select 6,sysdate -1-(1/(1440*60)) from dual)
select sysdate,a.* from rt a;
Row#     SYSDATE     SNO     SDATE
1     01-Aug-2007 2:32:06 PM     1     01-Aug-2007 2:32:06 PM
2     01-Aug-2007 2:32:06 PM     2     01-Aug-2007 2:32:05 PM
3     01-Aug-2007 2:32:06 PM     3     01-Aug-2007 2:31:06 PM
4     01-Aug-2007 2:32:06 PM     4     01-Aug-2007 2:02:06 PM
5     01-Aug-2007 2:32:06 PM     5     31-Jul-2007 2:32:07 PM
6     01-Aug-2007 2:32:06 PM     6     31-Jul-2007 2:32:05 PM
with rt as
(select 1 sno ,sysdate sdate from dual union all
select 2,sysdate - (1/(1440*60)) from dual union all
select 3,sysdate - (1/(1440)) from dual union all
select 4,sysdate - (30/(1440)) from dual union all
select 5,sysdate -1+(1/(1440*60)) from dual union all
select 6,sysdate -1-(1/(1440*60)) from dual)
select sysdate,a.* from rt a where sdate >= sysdate - 1;
Row#     SYSDATE     SNO     SDATE
1     01-Aug-2007 2:32:12 PM     1     01-Aug-2007 2:32:12 PM
2     01-Aug-2007 2:32:12 PM     2     01-Aug-2007 2:32:11 PM
3     01-Aug-2007 2:32:12 PM     3     01-Aug-2007 2:31:12 PM
4     01-Aug-2007 2:32:12 PM     4     01-Aug-2007 2:02:12 PM
5     01-Aug-2007 2:32:12 PM     5     31-Jul-2007 2:32:13 PM

Sql Query-Reply quickly ..pls

I have 3 tables:
CREATE TABLE MARKET (
MARKET_ID INTEGER NOT NULL,
MARKET_DEFINITION_ID     INTEGER NOT NULL,
MARKET_DATE     TIMESTAMP NOT NULL,
MARKET_HOUR     INTEGER,
MDS_UPDATE_DTS TIMESTAMP NOT NULL,
MDS_UPDATE_USER VARCHAR2(30) NOT NULL,
PRIMARY KEY (MARKET_ID),
CREATE TABLE EVENT_DEFINITION (
EVENT_DEFINITION_ID INTEGER NOT NULL,
EVENT_DESC VARCHAR2(255) NOT NULL,
DIRECTIVE VARCHAR2(30) NOT NULL,
EVENT_TIME_OFFSET INTEGER NOT NULL,
OFFSET_UNITS VARCHAR2(10) NOT NULL,
PRIMARY KEY (EVENT_DEFINITION_ID));
CREATE TABLE MARKET_EVENT_HISTORY (
MARKET_EVENT_HISTORY_ID INTEGER NOT NULL,
EVENT_DEFINITION_ID INTEGER NOT NULL,
MARKET_ID INTEGER NOT NULL,
MDS_INSERT_DTS TIMESTAMP NOT NULL,
PRIMARY KEY (MARKET_EVENT_HISTORY_ID),
FOREIGN KEY (MARKET_ID)REFERENCES MARKET,
FOREIGN KEY (EVENT_DEFINITION_ID) REFERENCES EVENT_DEFINITION);
COMMIT;
and the data in the :
MARKET_EVENT_HISTORY will be some thing like this:
MARKET_EVENT_HISTORY_ID MDS_INSERT_DTS
"1",                     "31-JUL-02 09.00.00.123000 AM"
"2",                     "30-JUL-07 08.00.00.000000 AM"
"3",                    "30-JUL-07 11.00.00.000000 PM"
"4",                    "31-JUL-07 12.00.00.000000 PM"
"5",                     "31-JUL-07 11.00.00.000000 PM"
"6",               "31-JUL-07 10.00.00.000000 PM"
"7",                     "31-JUL-07 07.00.00.000000 PM"
"8",                     "31-JUL-02 01.00.00.123000 AM"
"9",                     "30-JUL-02 01.00.00.123000 AM"
"10",                    "30-JUL-02 06.00.00.123000 AM"
"11",                    "31-JUL-02 04.00.00.123000 AM"
"21",                "31-JUL-02 03.00.00.123000 AM"
and assume rest of the columns is having some data.
I want 5 columns from the above 3 tables
From Market Table
MARKET_DATE     TIMESTAMP NOT NULL,
MARKET_HOUR     INTEGER,
From Event definiton table
OFFSET_UNITS VARCHAR2(10) NOT NULL
From Market event history table
MDS_INSERT_DTS
i want a query for all the 5 columns with the help of join .
But the conditon is :
Based upon MDS_INSERT_DTS i need to retreive the last 24 hours records in the data base.
Assume MDS_INSERT_DTS is sysdate.
Ex:
Suppose the sysdate is
31-JUL-07 10.00.00.000000 PM"
then the last 24 hours records available in the data base(MDS_INSERT_DTS ) should be retrieved like below.
                    "30-JUL-07 08.00.00.000000 AM"
                    "31-JUL-02 09.00.00.123000 AM"
                    etc all the records of the last 24 hours has to be retrevied .
we have the relation ship between the tables in Market_Event_History.
send me the reply immediately...plssssssssssssss..........urgent

with rt as
(select 1 sno ,sysdate sdate from dual union all
select 2,sysdate - (1/(1440*60)) from dual union all
select 3,sysdate - (1/(1440)) from dual union all
select 4,sysdate - (30/(1440)) from dual union all
select 5,sysdate -1+(1/(1440*60)) from dual union all
select 6,sysdate -1-(1/(1440*60)) from dual)
select sysdate,a.* from rt a;
Row#     SYSDATE     SNO     SDATE
1     01-Aug-2007 2:32:06 PM     1     01-Aug-2007 2:32:06 PM
2     01-Aug-2007 2:32:06 PM     2     01-Aug-2007 2:32:05 PM
3     01-Aug-2007 2:32:06 PM     3     01-Aug-2007 2:31:06 PM
4     01-Aug-2007 2:32:06 PM     4     01-Aug-2007 2:02:06 PM
5     01-Aug-2007 2:32:06 PM     5     31-Jul-2007 2:32:07 PM
6     01-Aug-2007 2:32:06 PM     6     31-Jul-2007 2:32:05 PM
with rt as
(select 1 sno ,sysdate sdate from dual union all
select 2,sysdate - (1/(1440*60)) from dual union all
select 3,sysdate - (1/(1440)) from dual union all
select 4,sysdate - (30/(1440)) from dual union all
select 5,sysdate -1+(1/(1440*60)) from dual union all
select 6,sysdate -1-(1/(1440*60)) from dual)
select sysdate,a.* from rt a where sdate >= sysdate - 1;
Row#     SYSDATE     SNO     SDATE
1     01-Aug-2007 2:32:12 PM     1     01-Aug-2007 2:32:12 PM
2     01-Aug-2007 2:32:12 PM     2     01-Aug-2007 2:32:11 PM
3     01-Aug-2007 2:32:12 PM     3     01-Aug-2007 2:31:12 PM
4     01-Aug-2007 2:32:12 PM     4     01-Aug-2007 2:02:12 PM
5     01-Aug-2007 2:32:12 PM     5     31-Jul-2007 2:32:13 PM

Very slow running SQL

Hello I have a stored procedure that I am using to load 5 tables from a stagin table. It reads each line and populates these tables. The staging table has 200K rows in it and I commit every 10,000 rows. Once released to production the staging table will have around 23 million rows in it to be processed. It is running very poorly. I am running Oracle 9.2.0.6. Here is the stored procedure. Any suggestions on how to make it faster. Is there a way to use bulk binding/insert and would that be better?
Thank you,
David
CREATE OR REPLACE procedure SP_LOAD_STAGED_CUST (runtime_minutes int)
is
staged_rec STAGE_CUSTOMER%ROWTYPE;
end_time          date     default sysdate + (nvl(runtime_minutes,1440)/1440);
test_row_id     number(38);
in_table_cnt     number(38);
BEGIN
-- POPULATE LOCATION AND ASSOCIATE TABLES AS NEEDED
insert into TMP_LOCATION (update_location_cd, country_cd, locale_cd)
select      distinct update_location_cd,
     country_cd,
     locale_cd
from stage_customer
update TMP_Location
set location_id = location_seq.nextval
insert /*+ APPEND */ into location
( location_id, location_cd, location_desc, short_description, sales_channel_id, location_type_id,
location_category_id, addr1, addr2, addr3, city, state_cd, postal_cd, country_cd, region_id,
timezone_id, locale_cd, currency_cd, phone_num, alt_phone_num, fax_num, alt_fax_num,
email_addr, alt_email_addr, is_default, create_modify_timestamp)
select
location_id,
update_location_cd,
null,
null,
fn_sales_channel_default(),
null,
null,
null, null, null,
null, null, null ,
nvl(country_cd, 'USA'),
null, null,
locale_cd,
'USD',
null, null, null, null, null, null,
0,
sysdate
from TMP_LOCATION
where update_location_cd not in (select location_cd from location);
commit
insert into TMP_ASSOCIATE (associate_number, update_location_cd)
select      associate_number, min(update_location_cd)
     from stage_customer
where associate_number is not null
group by associate_number;
update TMP_ASSOCIATE
set associate_id = associate_seq.nextval
insert /*+ APPEND */ into associate
select
associate_id ,
null,
associate_number,
(select nvl(location_id, fn_location_default())
     from location
     where location.location_cd = tmp_associate.update_location_cd)
from TMP_ASSOCIATE
where not exists (select associate_id from associate
          where associate.associate_number = tmp_associate.associate_number)
delete from tmp_associate;
commit;
insert into TMP_ASSOCIATE (associate_number, update_location_cd)
select      alt_associate_number, min(update_location_cd)
     from stage_customer
where alt_associate_number is not null
group by alt_associate_number
update TMP_ASSOCIATE
set associate_id = associate_seq.nextval
insert /*+ APPEND */ into associate
select
associate_id ,
null,
associate_number,
(select nvl(location_id, fn_location_default())
     from location
     where location.location_cd = tmp_associate.update_location_cd)
from TMP_ASSOCIATE
where not exists (select associate_id from associate
          where associate.associate_number = tmp_associate.associate_number);
commit;
select min(row_id) -1 into test_row_id from stage_customer ;
WHILE sysdate < end_time
LOOP
select *
into staged_rec
from Stage_Customer
where row_id = test_row_id + 1;
if staged_rec.row_id is null
then
COMMIT;
EXIT;
end if;
-- EXIT WHEN staged_rec.row_id is null;
-- INSERTS TO CUSTOMER TABLE (IN LOOP - DATA FROM STAGE CUSTOMER TABLE)
insert /*+ APPEND */ into customer (
     customer_id,
     customer_acct_num,
     account_type_id,
     acct_status,
     discount_percent,
     discount_code,
     uses_purch_order,
     business_name,
     tax_exempt_prompt,
     is_deleted,
     name_prefix,
     first_name,
     middle_name,
     last_name,
     name_suffix,
     nick_name,
     alt_first_name,
     alt_last_name,
     marketing_source_id,
     country_cd,
     locale_cd,
     email_addr,
     email_addr_valid,
     alt_email_addr,
     alt_email_addr_valid,
     birth_date,
     acquisition_sales_channel_id,
     acquisition_location_id,
     home_location_id,
     salesperson_id,
     alt_salesperson_id,
     customer_login_name,
     age_range_id,
     demographic_role_id,
     education_level_id,
     gender_id,
     household_count_id,
     housing_type_id,
     income_range_id,
     lifecycle_type_id,
     lifetime_value_score_id,
     marital_status_id,
     religious_affil_id
values (
     staged_rec.row_id,
     staged_rec.customer_acct_num,
     nvl(staged_rec.account_type_id, fn_account_type_default()),
     1,
     staged_rec.discount_percent,
     staged_rec.discount_cd,
     staged_rec.pos_allow_purchase_order_flag,
     staged_rec.business_name,
     staged_rec.pos_tax_prompt,
     staged_rec.is_deleted,
     staged_rec.name_prefix,
     staged_rec.first_name,
     staged_rec.middle_name,
     staged_rec.last_name,
     staged_rec.name_suffix,
     staged_rec.nick_name,
     staged_rec.alt_first_name,
     staged_rec.alt_last_name,
     staged_rec.new_marketing_source_id,
     staged_rec.country_cd,
     staged_rec.locale_cd,
     staged_rec.email_addr,
     nvl2(staged_rec.email_addr,1,0),
     staged_rec.alt_email_addr,
     nvl2(staged_rec.alt_email_addr,1,0),
     staged_rec.birth_date,
     staged_rec.SALES_CHANNEL_ID,
     (select location_id from location where location_cd = staged_rec.update_location_cd),
     (select location_id from location where location_cd = staged_rec.update_location_cd),
     (select min(a.associate_id) from associate a
          where staged_rec.associate_number = a.associate_number
          and a.location_id =
               (select location_id from location where location_cd = staged_rec.update_location_cd)),
     (select min(a.associate_id) from associate a
          where staged_rec.alt_associate_number = a.associate_number
          and a.location_id =
               (select location_id from location where location_cd = staged_rec.update_location_cd)),
     staged_rec.customer_login_name,
     fn_age_range_default(),
     fn_demographic_role_default(),
     fn_education_level_default(),
     fn_gender_default(),
     fn_household_cnt_default(),
     fn_housing_type_default(),
     fn_income_range_default(),
     fn_lifecycle_type_default(),
     fn_lifetime_val_score_default(),
     fn_marital_status_default(),
     fn_religious_affil_default()
-- INSERTS TO PHONE TABLE ( IN LOOP -DATA FROM STAGE CUSTOMER TABLE)
if staged_rec.home_phone is not null
then
insert /*+ APPEND */ into phone ( customer_id, phone_type_id, phone_num, is_valid, is_primary)
     values (staged_rec.row_id, 1, staged_rec.home_phone, 1, 1);
end if;
if staged_rec.work_phone is not null
then
insert /*+ APPEND */ into phone (customer_id, phone_type_id, phone_num, is_valid,is_primary)
     values (staged_rec.row_id, 2, staged_rec.work_phone, 1, 0);
end if;
if staged_rec.mobile_phone is not null
then
insert /*+ APPEND */ into phone ( customer_id, phone_type_id, phone_num, is_valid, is_primary)
     values (staged_rec.row_id, 3, staged_rec.home_phone, 1, 0);
end if;
if staged_rec.work_phone is not null
then
insert /*+ APPEND */ into phone (customer_id, phone_type_id, phone_num, is_valid,is_primary)
     values (staged_rec.row_id, 4, staged_rec.work_phone, 1, 0);
end if;
-- INSERTS TO CUSTOMER ADDR TABLE ( IN LOOP - DATA FROM STAGE CUSTOMER TABLE)
if staged_rec.address1 is not null
then
insert /*+ APPEND */ into customer_addr (customer_id, address_type_id, address1, address2, address3, city, state_cd, postal_cd, country_cd, region_id,
               is_primary, is_valid, cannot_standardize, is_standardized)
     values (staged_rec.row_id, fn_address_type_default(), staged_rec.address1, staged_rec.address2, staged_rec.address3,
          staged_rec.city, staged_rec.state_cd, staged_rec.postal_cd, staged_rec.country_cd, staged_rec.region, 1, 1, 0, 0);
end if;
-- INSERTS TO CUSTOMER_STATE_TAX_ID
if staged_rec.pos_default_tax_id is not null
then
insert /*+ APPEND */ into customer_state_tax_id (state_cd, customer_id, tax_id, expiration_date)
     values (staged_rec.state_cd, staged_rec.row_id, staged_rec.pos_default_tax_id,
          staged_rec.pos_tax_id_expiration_date);
end if;
-- REMOVE STAGE CUSTOMER ROW (IN LOOP - DELETE CUSTOMER FROM STAGE_CUSTOMER TABLE)
delete from stage_customer
where row_id = staged_rec.row_id;
-- COMMIT AFTER EVERY 10,000 CUSTOMERS
if mod(staged_rec.row_id, 100) = 0
then
commit;
end if;
-- INCREMENT ROW ID TO BE RETRIEVED
test_row_id := test_row_id + 1;
END LOOP;
EXCEPTION
     WHEN NO_DATA_FOUND
     THEN
     COMMIT;
END;
Message was edited by:
JesusLuvR
Message was edited by:
JesusLuvR

You want to do as much processing as you can in single large sql statements, not row by row. You also want to do as few statemetns as possible. Most of what you are doing looks to me like it could be done as a series of single sql statements. For example, the three statements that populate location ca nbe condensed to a single statement like:
INSERT /*+ APPEND */ INTO location
   (location_id, location_cd, location_desc, short_description, sales_channel_id,
    location_type_id, location_category_id, addr1, addr2, addr3, city, state_cd,
    postal_cd, country_cd, region_id, timezone_id, locale_cd, currency_cd,
    phone_num, alt_phone_num, fax_num, alt_fax_num, email_addr, alt_email_addr,
    is_default, create_modify_timestamp)
SELECT location_seq.NEXTVAL, update_location_cd, NULL, NULL,
       fn_sales_channel_default(), NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,
       NVL(country_cd, 'USA'), NULL, NULL, locale_cd, 'USD', NULL, NULL, NULL,
       NULL, NULL, NULL, 0, sysdate
FROM (SELECT DISTINCT update_location_cd, country_cd, locale_cd
      FROM stage_customer
      WHERE update_location_cd NOT IN (SELECT location_cd FROM location);You can easily do a similar change to the statemetns populating associates.
I don't, off hand, see anything in the intert into customers that seems to require row by row processing. Just do it in a large insert. I would also be tempted to replace the many values of the form (select location_id from location where location_cd = staged_rec.update_location_cd) to a join with location. As far as I can see, you would only need one copy of location.
I also notince that you have a number of function calls in the values list. I would look carefully at whether you need them at all, since they seem to have no parameters, so should return the same value every time. Minimally, I would consider calling them once each, storing the values in variables and use those variables in the insert.
HTH
John

Question about DBMS_JOB

Hi , I have this job that I would like to modify so that it runs only from sunday to thursday and starts at 8:00 till 14:30.
DECLARE
X NUMBER;
BEGIN
SYS.DBMS_JOB.SUBMIT
    ( job       => X
     ,what      => 'update_port_prices;'
     ,next_date => sysdate
     ,interval => ' SYSDATE + 0.010416666666666666666666666666667 /* every quarter an hour */ '
     ,no_parse => TRUE
SYS.DBMS_OUTPUT.PUT_LINE('Job Number is: ' || to_char(x));
END;
commit;

Try this
create or replace function next_execution return date as
v_dt date;
begin
v_dt:= case when to_char(sysdate,'HH24MI') between 0800 and 1415 then
trunc(sysdate,'MI')+15/1440
when to_char(sysdate,'DY') in ('SUN','MON','TUE','WED') then
trunc(sysdate)+1+8/24
else next_day(trunc(sysdate),'SUN')+8/24 end;
return v_dt;
end;
--Checking function output
PRAZY@11gR1> select to_char(sysdate,'MM/DD/YYYY HH24:MI:SS') from dual;
TO_CHAR(SYSDATE,'MM
03/16/2010 16:46:00
Elapsed: 00:00:00.00
PRAZY@11gR1> select to_char(next_execution,'MM/DD/YYYY HH24:MI:SS') from dual;
TO_CHAR(NEXT_EXECUT
03/17/2010 08:00:00
Elapsed: 00:00:00.00
--job scheduling
declare
v_num number;
begin
dbms_job.submit(v_num,'update_port_prices;',sysdate,'next_execution()');
dbms_output.put_line('Job Number :'||v_num);
end;Regards,
Prazy
Edited by: Prazy on Mar 16, 2010 4:42 PM
Removed THU as it has to run only from sunday to thursday & added sysdate in function test & fixed some typo's ;)

Combine Two table columns

SELECT providerid, SUM(COUNT) FROM appdev.usercounts
WHERE counttime BETWEEN SYSDATE - 30/1440 AND SYSDATE- 15/1440 GROUP BY providerid ORDER BY providerid
W     697
U     813
T     143
S     2
SELECT providerid, SUM(COUNT) FROM appdev.usercounts
WHERE counttime BETWEEN (SYSDATE -1) - 30/1440 AND (SYSDATE-1)- 15/1440 GROUP BY providerid ORDER BY providerid;
W     450
U     571
T     80
S     2
I wnat output as like
W     697 450
U     813 571
T     143 80
S     2 2
please help
Thanks
Praveen

"COUNT", is that the real column name? It could become very confusing.
Not tested for obvious reasons, but this should be close :
SELECT providerid
     , sum(
         case when counttime between sysdate - 30/1440 and sysdate - 15/1440 then count end
       ) as cnt1
     , sum(
         case when counttime between (sysdate-1) - 30/1440 and (sysdate-1) - 15/1440 then count end
       ) as cnt2
FROM appdev.usercounts
WHERE ( counttime between sysdate - 30/1440 and sysdate - 15/1440 )
   OR ( counttime between (sysdate-1) - 30/1440 and (sysdate-1) - 15/1440 )
GROUP BY providerid
ORDER BY providerid
;

Check Sysdate for a specified Time

Hi,
I have requirement to check sysdate for particular time in a script,If it has systime is equal to that time in the evening, I need to exit the script and write messages to log file. I am doing the following but it's not working properly, Can anybody help with this?
if sysdate > to_date('19:20','HH24:MI')then
v_timeout := 'Y';
v_output := 'Database Shutdown time... Exiting Script';
utl_file.put_line(v_file_handle, v_output);
Thanks
Kalyan
exit;
end if;

to_date('19:20','HH24:MI')The date out going by this command is the first day of the current month...
SQL> select sysdate, to_date('19:20','HH24:MI') from dual;
SYSDATE TO_DATE('19:20','HH
14-04-2008 13:19:42 01-04-2008 19:20:00You have to use the current day instead :
SQL> select sysdate, trunc(sysdate)+19/24+20/1440 from dual;
SYSDATE TRUNC(SYSDATE)+19/2
14-04-2008 13:21:21 14-04-2008 19:20:00Nicolas.
Paolo, how did you made ? I have a opened window without your response above mine, and an other with your post... hmmm some strange forum behaviour...
Message was edited by:
N. Gasparotto

Using SYSDATE in RMAN script

Hi
Each night I move an RMAN backup from a PROD server to a test server.
Then using RMAN script with NOCATALOG, I restore/recover the database on the test server as show below:
RUN{
ALLOCATE CHANNEL dev1 DEVICE TYPE DISK;
STARTUP NOMOUNT;
RESTORE CONTROLFILE from AUTOBACKUP;
SHUTDOWN;
STARTUP FORCE MOUNT;
RESTORE DATABASE;
RECOVER DATABASE;
ALTER DATABASE OPEN RESETLOGS;
Of course the script fails at the "RECOVER DATABASE" command because it runs out of ARCH logs to apply at some point and therefore never gets to the final command of DATABASE OPEN. I would like to change that line to something like:
RECOVER DATABASE UNTIL TIME 'SYSDATE:02:45:00';
The above fails with an RMAN error of "expected number".
Of course I could use SET UNTIL TIME "to_date('SYSDATE 02:45:00','DD-MON-YYYY HH24:MI:SS')"; although I have the same problem, i.e., getting RMAN to accept SYSDATE as the date portion of the point in time recovery.
Oddly, from an RMAN prompt and within brackets, I can issue the following command successfully:
RMAN {
set until time "to_date('SYSDATE 02:45:00','DD-MON-YYYY HH24:MI:SS')";
command completed successfully
So my question is, how can I use SYSDATE in an RMAN point in time recovery and have sysdate represent the current date, either in the RECOVER DATABASE line or the SET UNTIL TIME line?
Thanks.

user522620 wrote:
Hi
Each night I move an RMAN backup from a PROD server to a test server.
Then using RMAN script with NOCATALOG, I restore/recover the database on the test server as show below:
RUN{
ALLOCATE CHANNEL dev1 DEVICE TYPE DISK;
STARTUP NOMOUNT;
RESTORE CONTROLFILE from AUTOBACKUP;
SHUTDOWN;
STARTUP FORCE MOUNT;
RESTORE DATABASE;
RECOVER DATABASE;
ALTER DATABASE OPEN RESETLOGS;
So my question is, how can I use SYSDATE in an RMAN point in time recovery and have sysdate represent the current date, either in the RECOVER DATABASE line or the SET UNTIL TIME line?
Thanks.
Given:
SQL> alter session set NLS_DATE_FORMAT='DD-MON-YYYY HH24:MI:SS';
SQL> select to_char(trunc(sysdate)+(2/24)+(45/1440),'dd-mon-yyyy hh24:mi') from
dual;
TO_CHAR(TRUNC(SYS
25-JUN-2012 02:45:00
Use this (uses to_date)
RUN{
set until time "to_date(trunc(sysdate)+(2/24)+(45/1440),'DD-MON-YYYY HH24:MI:SS')";
ALLOCATE CHANNEL dev1 DEVICE TYPE DISK;
STARTUP NOMOUNT;
RESTORE CONTROLFILE from AUTOBACKUP;
SHUTDOWN;
STARTUP FORCE MOUNT;
RESTORE DATABASE;
RECOVER DATABASE;
ALTER DATABASE OPEN RESETLOGS;
## if that command fails, it will fall to the next command as the run blocks are autonomous.
RUN{
ALTER DATABASE OPEN RESETLOGS;
}

Fetch records under 20 seconds in my view with sysdate

hi all
I want a query in my view that it returns records under 20 seconds and
after 20 seconds my query do not show any records
my query is :
select * from cnfenterexit
where cnf06date > sysdate - (1/86400000)
but it didn't work.
when I used following query whit this numbers ,
it is OK and show records under 56 seconds
select * from cnfenterexit
where cnf06date > sysdate - (3/1440)
please explain how to get under 20 seconds in me query?

or
select *
from cnfenterexit
where cnf06date > sysdate -TO_DSINTERVAL('0 00:0:20')TO_DSINTERVAL('0 00:0:20') = 0 days, 0 hours, 0 minutes, 20 seconds
Timo

Query help : Query to get values SYSDATE-1 18:00 hrs to SYSDATE 08:00 hrs

Hi Team
I want the SQl query to get the data for the following comparison : -
Order Created is a Date Column , and i want to find out all the values from (SYSDATE-1) 18:00 hours to SYSDATE 08:00 hours
i.e.
(SYSDATE-1) 18:00:00 < Order.Created < SYSDATE 08:00:00.
Regards

Hi, Rohit,
942281 wrote:
If i want the data in the below way i.e.
from (SYSDATE-1) 18:00 hours to SYSDATE 17:59 hours ---> (SYSDATE-1) 18:00:00 < Order.Created < SYSDATE 07:59:00.If you want to include rows from exactly 18:00:00 yesterday (but no earlier), and exclude rows from exatly 08:00:00 today (or later), then use:
WHERE   ord_dtl.submit_dt >= TRUNC (SYSDATE) - (6 / 24)
AND     ord_dtl.submit_dt < TRUNC (SYSDATE) + (8 / 24)
So can i use the below format : -
ord_dtl.submit_dt BETWEEN trunc(sysdate)-(6/24) and trunc(sysdate)+(7.59/24) . Please suggest . .59 hours is .59 * 60 * 60 = 2124 seconds (or .59 * 60 = 35.4 minutes), so the last time included in the range above is 07:35:24, not 07:59:59.
If you really, really want to use BETWEEN (which includes both end points), then you could do it with date arithmentic:
WHERE   ord_dtl.submit_dt BETWEEN TRUNC (SYSDATE) - (6 / 24)
                  AND         TRUNC (SYSDATE) + (8 / 24)
                                           - (1 / (24 * 60 * 60))but it would be simpler and less error prone to use INTERVALs, as Karthick suggested earlier:
WHERE   ord_dtl.submit_dt BETWEEN TRUNC (SYSDATE) - INTERVAL '6' HOUR
                  AND         TRUNC (SYSDATE) + INTERVAL '8' HOUR
                                           - INTERVAL '1' SECONDEdited by: Frank Kulash on Apr 17, 2013 9:36 AM
Edited by: Frank Kulash on Apr 17, 2013 11:56 AM
Changed "- (8 /24)" to "+ (8 /24)" in first code fragment (after Blushadown, below)

How to insert the sysdate time into the database

hi all,
when i execute the following query,
insert into table_name
(date_field)
values
(to_date(sysdate, 'yyyy/mm/dd hh24:mi:ss'));
The value is inserted as 08-02-12 12:00:00. In this query, it always stores the default time as 12 a.m.
But i need to insert the original system time not the default one.
please help me how to rectify it.

I do not understand as to why you are using the to_date function on sysdate since sysdate is already in date format.
If date_field is of the data type date, then the following dml should work.
insert into table_name(date_field) values (sysdate);
In case date_field is a varchar2 field and you want to store the date and time in a string format, then the following statement should work.
insert into table_name (date_field)
values (to_char(sysdate, 'yyyy/mm/dd hh24:mi:ss'));

SYSDATE+(60*24)/1440

Similar Messages

Maybe you are looking for