I am doing comparison between several dates in my stored procedure.
TODAY := TO_DATE(TO_CHAR(SYSDATE, 'DD-MON-YYYY') ||
' 09:00:00', 'DD-MON-YYYY HH24:MI:SS');
IF PREVIOUS_DATE < TODAY AND
TO_DATE(CURRENT_DATE, 'DD-MON-YYYY HH24:MI:SS') >= TODAY THEN
-- do something
ELSE
-- do something else
When I set CURRENT_DATE = SYSDATE, it did not get into the IF part. Can someone tell me where did i do wrong?
"CURRENT_DATE is of type VARCHAR2 "
Well that scuppers my first idea 8-)
But I think the problem is in your use of CURRENT_DATE. Assuming that is the Oracle function there are two potential problems:
CURRENT_DATE is a DATE datatype and so has a time element. There's no point in casting it to a date.
CURRENT_DATE returns a value adjusted for the system timezone. Which is presumably the point of your IF test. But do you know what the values are?
But it remains a straightforward debugging task. Check your values to understand what's happening. Here is an example using DBMS_OUTPUT (AKA The Devil's Debugger) because I don't know if you're working in an environment with better tools.
TODAY := trunc(sysdate) + 9/24;
dbms_output.put_line('TODAY = '||to_char(today, 'DD-MON-YYYY HH24:MI:SS'));
dbms_output.put_line('PREVIOUS_DATE = '||to_char(previous_date, 'DD-MON-YYYY HH24:MI:SS'));
dbms_output.put_line('CURRENT_DATE = '||current_date);
IF PREVIOUS_DATE < TODAY AND
to_date(CURRENT_DATE, 'DD-MON-YYYY HH24:MI:SS') >= TODAY
THEN
....
Might as well simplify your code at the same time.
Remember, to see output from DBMS_OUTPUT.PUT_LINE you need to SET SERVEROUTPUT ON for whatever client you're using.
Incidentally, it is bad practice to declare variables which share the same name as Oracle built-ins. This is why it is a good idea to add a scoping prefix (l_ for local, p_ for parameter, etc) to our declarations. We don't have to do the full Hungarian to get a lot of benefit.
Related
Let me start with I have a confused Oracle table that has 2 particular columns in it, 1 for issuedate VARCHAR2(10) and one for compdate VARCHAR2(8). The NLS_DATE_FORMAT for the system session is 'YYYY-MM-DD HH24:MI:SS'. I cannot change the NLS_DATE_FORMAT as there are a significant set of SELECT's that use this format to convert other timestamps into dates.
issuedate is 'MMDDYYYY' COMPDATE is 'MM/DD/YYYY'
here are the portions of the Select in question
co.issuedate issued,
to_date(substr(ae.cdts, 1,8)) DATE_JOB_OPENED,
TO_DATE(substr(ae.xdts, 1,8)) DATE_JOB_CLOSED,
co.compdate WORKED,
The goal is to subtract
issued-WORKED
and get the result in number of days.
Guidance is appreciated
The first part of the solution should be to fix your table so that you are storing date values in a DATE data type and NOT in a VARCHAR2.
The second part of the solution should be to go back through all your old code and fix any instances where you use TO_DATE or TO_CHAR to make sure they always are passed a second argument to explicitly set the format model and to never rely on an implicit format model set by the NLS_DATE_FORMAT session parameter.
Since you are storing it as a VARCHAR2, just use TO_DATE and subtract to get the difference:
SELECT co.issuedate issued,
TO_DATE(substr(ae.cdts, 1,8), 'YYYYMMDD') DATE_JOB_OPENED,
TO_DATE(substr(ae.xdts, 1,8), 'YYYYMMDD') DATE_JOB_CLOSED,
co.compdate WORKED,
TO_DATE(co.issuedate, 'MMDDYYYY') - TO_DATE(co.compdate, 'MM/DD/YYYY')
AS days_difference
FROM ...;
and get the result in mm/dd/yyyy format.
If you really want the difference in days, months and years then it gets more complicated:
SELECT issued,
DATE_JOB_OPENED,
DATE_JOB_CLOSED,
worked,
TO_CHAR(
EXTRACT(MONTH FROM (issued_date - comp_date) YEAR TO MONTH),
'FM00'
)
||'/'||
TO_CHAR(
issued_date - ADD_MONTHS(comp_date, FLOOR(MONTHS_BETWEEN(issued_date, comp_date))),
'FM00'
)
||'/'||
TO_CHAR(
EXTRACT(YEAR FROM (issued_date - comp_date) YEAR TO MONTH),
'FM0000'
) AS difference
FROM (
SELECT co.issuedate issued,
TO_DATE(substr(ae.cdts, 1,8), 'YYYYMMDD') DATE_JOB_OPENED,
TO_DATE(substr(ae.xdts, 1,8), 'YYYYMMDD') DATE_JOB_CLOSED,
co.compdate WORKED,
TO_DATE(co.issuedate, 'MMDDYYYY') AS issued_date,
TO_DATE(co.compdate, 'MM/DD/YYYY') AS comp_date
FROM table_name co
-- ...
);
Which, for the sample data:
CREATE TABLE table_name (issuedate, compdate) AS
SELECT '10292021', '01/01/2021' FROM DUAL;
Outputs for the co related columns:
ISSUED
WORKED
DIFFERENCE
10292021
01/01/2021
10/28/0000
db<>fiddle here
I have two tables: SKN_ENJIN, and SKN_ENJIN_COUNT
SKN_ENJIN keeps track of usernames and emails.
SKN_ENJIN_COUNT is being used to populate a chart for a dashboard report.
I created this trigger earlier today:
create or replace trigger "BI_SKN_ENJIN_COUNT_TG"
after insert or update or delete on "SKN_ENJIN"
DECLARE
mCount NUMBER;
mDate DATE;
begin
select COUNT(ID) into mCount from SKN_ENJIN where Status = 1;
select TO_DATE(CURRENT_DATE, 'DD-MM-YYYY') into mDate from dual;
MERGE INTO SKN_ENJIN_COUNT c
USING dual d
ON (c.Count_date = mDate)
WHEN MATCHED THEN
UPDATE SET c.Member_count = mCount
WHEN NOT MATCHED THEN
INSERT (Count_date, Member_count)
VALUES (mDate, mCount);
end;
Up until about 10 minutes ago, the trigger worked beautifully. Suddenly, the trigger started throwing ORA-01843: not a valid month
I have tried changing CURRENT_DATE to SYSDATE(), I have tried changing TO_DATE to TO_CHAR. These approaches seemed to cause more errors to appear. What am I missing, and what should I change to solve this problem?
There's no need to call TO_DATE on CURRENT_DATE or SYSDATE. These functions already return DATEs, so there's no need to do any conversion.
In fact, calling TO_DATE on something that is already a DATE forces Oracle to convert it to a string using NLS settings (NLS_DATE_FORMAT) and the convert it back to a date using a given date format picture. If the date format picture you are using does not match NLS_DATE_FORMAT, you will likely end up with errors or incorrect values.
Instead of writing
select TO_DATE(CURRENT_DATE, 'DD-MM-YYYY') into mDate from dual;
you can write
select CURRENT_DATE into mDate from dual;
or just
mDate := CURRENT_DATE;
I don't know what type of the Count_date column in your SKN_ENJIN_COUNT table is, but if it is DATE, it is similarly incorrect to call TO_DATE on it.
I think I found a solution while stumbling my way through it
. It seems that the format of the date is extremely important. Earlier my formatting had been DD-MM-YYYY, when I used MM-DD-YYYY and just a touch of refactoring (ON (TO_DATE(c.Count_date, 'MM-DD-YYYY') = TO_DATE(mDate, 'MM-DD-YYYY')) the script worked without a fuss.
select COUNT(ID) into mCount from SKN_ENJIN where Status = 1;
select sysdate into mDate from dual;
MERGE INTO SKN_ENJIN_COUNT c
USING dual d
ON (TO_DATE(c.Count_date, 'MM-DD-YYYY') = TO_DATE(mDate, 'MM-DD-YYYY'))
WHEN MATCHED THEN
UPDATE SET c.Member_count = mCount
I´m trying to calculate the difference between two dates in Oracle and getting the result as a TimeStamp. This is the easiest thing to do in SQL Server, but it seems that Oracle does not have a easy way to solve this. I refuse to believe that I have to write that much code to get what I need. Can someone tell me if there is a easier way to get that difference?:
SELECT TO_CHAR(EXTRACT(HOUR FROM NUMTODSINTERVAL(enddate-startdate, 'DAY')), 'FM00')
|| ':' ||
TO_CHAR(EXTRACT(MINUTE FROM NUMTODSINTERVAL(enddate-startdate, 'DAY')), 'FM00')
|| ':' ||
TO_CHAR(EXTRACT(SECOND FROM NUMTODSINTERVAL(enddate-startdate, 'DAY')), 'FM00')
I need the result be something like:
enddate = '2017-03-01 17:30:00'
startdate = '2017-03-01 10:00:00'
difference: 07:30:00
Substract the two dates. Add the result to the current date (without any time component, trunc(sysdate)) and show only the time.
select to_char(trunc(sysdate) + (to_date('2017-03-01 17:30:00', 'YYYY-MM-DD HH24:MI:SS') -
to_date('2017-03-01 10:00:00', 'YYYY-MM-DD HH24:MI:SS'))
,'HH24:MI:SS')
from dual
I've new to PL/SQL (and it's been a while since I've used vanilla SQL). I've got a query that I inherited that I'm trying to schedule in TOAD. In order for that to work I have to change hard coded date references to be calculated at run time.
To that end I added a Declare statement to the front of the query, added the necessary constants, setting them at declaration, and then had the query use them.
When I try to execute an error gets thrown saying a Select Into. To my understanding, SELECT Into is used to set a variable based on a value in the db (based on Constants in Oracle SQL query), whereas I'm looking to define the value independent of any value in the db (in this case the date on the server). The full error follows:
ORA-06550: line 6, column 5:
PLS-00428: an INTO clause is expected in the Select statement
So I'm looking for a little guidance on where my understanding of variables/constants in PL/SQL is off, and also help with getting the following to execute:
DECLARE OLD CONSTANT char(11):= to_Char(SYSDATE - 6, 'DD-MON-YYYY');
NEW CONSTANT char(11):= to_char(SYSDATE, 'DD-MON-YYYY');
BEGIN
SELECT CASE
WHEN (userhost LIKE 'a%'
AND userid IN ('s',
'sub')) THEN 'BATCH'
WHEN userid LIKE 'N%' THEN 'N'
WHEN ((userhost LIKE 'b%'
OR userhost LIKE 'c%')
AND userid IN ('s',
'sub')) THEN 'Forms'
WHEN ((userid LIKE '%_IU%'
OR userid LIKE 'RPT%'
OR userid IN ('q',
'r',
'p'))
AND userhost <> 'n%') THEN 'Interface'
ELSE 'Other'
END app_type , round(sum(sessioncpu/100), 1) cpu_seconds , (sum(sessioncpu/100) / (119*1*60*60) * 100) pct_of_cpu,
trunc(ntimestamp#,'MI')
FROM PERFSTAT.AUD$_ARCHIVE
WHERE ntimestamp# BETWEEN to_timestamp(OLD || ' 23:59','DD-MON-YYYY HH24:MI') AND to_timestamp(NEW || ' 00:00','DD-MON-YYYY HH24:MI')
AND logoff$time < to_date(NEW || ' 00:00','DD-MON-YYYY HH24:MI')
GROUP BY CASE
WHEN (userhost LIKE 'a%'
AND userid IN ('s',
'sub')) THEN 'BATCH'
WHEN userid LIKE 'N%' THEN 'N'
WHEN ((userhost LIKE 'b%'
OR userhost LIKE 'c%')
AND userid IN ('s',
'sub')) THEN 'Forms'
WHEN ((userid LIKE '%_IU%'
OR userid LIKE 'RPT%'
OR userid IN ('q',
'r',
'p'))
AND userhost <> 'n%') THEN 'Interface'
ELSE 'Other'
END app_type,
trunc(ntimestamp#,'MI')
ORDER BY trunc(ntimestamp#,'MI'),
1;
END;
You have two issues here. The first is trying to use the CHAR datatype and then not giving it a length. This defaults to a CHAR(1), i.e. a single character. For memory concerns, you might also consider VARCHAR2 instead. https://docs.oracle.com/cd/E17952_01/refman-5.1-en/char.html
The second issue has to do with the INTO clause as mentioned in your question. When you run a SELECT statement in PL/SQL (not associated to DML), you have to give Oracle something to return the result set into. You can then use those variables, whether printing them, storing them, or doing processing with them.
I'd have to see the error, but I think that it may want you to set a length to your char. So, something like char(30).
Also, I'm a big fan of varchar2. Only uses as much space in the DB as the characters in the variable. So, it it's varchar2(500) and has 8 characters, it only uses 8 chars worth of memory.
Your query has an inherent flaw, in that anything that occurs between 23:59 and 0:00 will satisfy conditions at both ends of the range (e.g. something that happens at 23:59:30). If this query were my responsibility, I'd get rid of the variables and the text conversions altogether:
WHERE ntimestamp# >= TRUNC (SYSDATE) - 6
AND ntimestamp# < TRUNC (SYSDATE)
AND logoff$time < TRUNC (SYSDATE)
Using >= and < for dates where you want to avoid an overlap tends to be safer that using between.
Taking a closer look, I'm not sure what the point of your query using one minute before midnight on the lower bound is. That kind of thing is more typically done on the upper bound. Assuming that you're actually doing that for a reason, you can still get around transforming to a string by using either of the following:
WHERE ntimestamp# BETWEEN TRUNC (SYSDATE) - 6 - (1 / 24 / 60)
AND TRUNC (SYSDATE)
AND logoff$time < TRUNC (SYSDATE)
WHERE ntimestamp# BETWEEN TRUNC (SYSDATE)
- NUMTODSINTERVAL (6, 'DAY')
- NUMTODSINTERVAL (1, 'MINUTE')
AND TRUNC (SYSDATE)
AND logoff$time < TRUNC (SYSDATE)
All of that is really just an aside to your main problem though: you need to tell the interpreter what to do with the result of the query. That means that you need to provide a variable to put the result in, then (presumably) do something with the result. One way to do this is to use a cursor loop:
DECLARE
CURSOR cur_query IS
[your query goes here];
BEGIN
FOR r_query IN cur_query LOOP
DBMS_OUTPUT.put_line (r_query.app_type);
DBMS_OUTPUT.put_line (r_query.cpu_seconds);
DBMS_OUTPUT.put_line (r_query.pct_of_cpu);
END LOOP;
END;
Of course, the alternative is just to run your query as SQL, rather than PL/SQL. With the variables eliminated, that will be easier.
Comment Response
PL/SQL blocks are not intended to return query results, like you get if you run straight SQL in Toad. There are ways to fake it via functions that return user-defined types or pipelined functions, but you're better off writing SQL if you are able to (and, in this case, you should be able to).
I'm not sure what you mean by "the variables are supposed to dynamically set the date range to look at". The code provided is returning data relative to sysdate, not getting outside data. You can do that in the query as easily as you can in a PL/SQL block.
Sory, I have Question ? Why not show up when I execute some of its field content.
Please help me to fix it. Thank U
SELECT * FROM T_TRANS WHERE TIME_START = to_date('01-09-2014', 'DD-MM-YY');
Oracle is being a bit lenient with you with the date format mask, but you should use YYYY to make it clearer. But you're still providing a single point in time as midnight on that date:
select to_char(to_date('01-09-2014', 'DD-MM-YY'), 'YYYY-MM-DD HH24:MI:SS') as two_digit,
to_char(to_date('01-09-2014', 'DD-MM-YYYY'), 'YYYY-MM-DD HH24:MI:SS') as four_digit
from dual;
TWO_DIGIT FOUR_DIGIT
------------------- -------------------
2014-09-01 00:00:00 2014-09-01 00:00:00
Given the name of the column it's reasonable to assume you have other times, and your query won't match anything except exactly midnight. To find all records on that day, you need to provide a range:
SELECT * FROM T_TRANS
WHERE TIME_START >= to_date('01-09-2014', 'DD-MM-YYYY');
AND TIME_START < to_date('02-09-2014', 'DD-MM-YYYY');
... though I prefer the ANSI date notation for this sort of this:
WHERE TIME_START >= DATE '2014-09-01'
AND TIME_START < DATE '2014-09-02'
You could specify 23:59:59 on the same date, but that will break subtly when you use a timestamp field rather than a date field.
You could also truncate the value from the table (as #San said in a comment), or convert to a string for comparison (as #yammy showed in an answer), but either of those will prevent any index on the time_start column being used, affecting performance.
After reading Alex Poole suggestion, you should try:
SELECT * FROM T_TRANS WHERE to_char(TIME_START, 'DD-MM-YYYY') = '01-09-2014';
Shouldn't the format mask be "DD-MM-YYYY"?