Recently there is an issue from my side.
In the table in redshift I have 2 columns:
The visit_time, which mentions the start time of the session and in YYYY-MM-DD HH:MM:SS format; Another column which is the time_spent, and it stands for the millesecond that user spends in certain page and currently it is in float(8)
What I want to do is to Add the visit_time with time_spent in second and convert it into YYYY-MM-DD HH:MM:SS, for example 2018-09-20 21:00:55 + 2 = 2018-09-20 21:00:57, so that I can get the visit_end_time. I tried to add it with date add function like this
Select
dateadd(SECOND,CAST (a.timeonpage AS DECIMAL)/1000 ,a.visit_time::date) time_left,
But it returns error: Invalid operation: function pg_catalog.date_add("unknown", double precision, date) does not exist
To do this, I tried to used the dateadd function like this:
Select
dateadd(SECOND,CAST (a.timeonpage AS INTEGER)/1000 ,a.visit_time::date) time_left,
It does not return error, but it returns the start of the date e.g 2018-09-23 00:00:00, which is not sth that I need.
What is the error that I made and how can I solve it?
Many thanks!
The dateadd function does require an integer for the interval. The problem with your second example is that when you cast a.visit_time to a date type it truncates it to the start of that day (removes the time component). Cast it to a timestamp instead:
select dateadd(second, round(a.timeonpage/1000.0)::integer, a.visit_time::timestamp);
select dateadd(second, round(2345/1000.0)::integer, '2018-09-20 21:00:55'::timestamp);
date_add
---------------------
2018-09-20 21:00:57
Also if you are storing visit_time as a string I would recommend making the column type a timestamp.
Related
There is a problem with a query I use to report.I get an error comparing a value stored as a timestamp with data saved yesterday.
query:
SELECT * FROM PIECE P, PIECE_ATTRB PA WHERE P.PIECE_NUM_ID=PA.PIECE_NUM_ID
AND PA.ATTRB_CODE='PRODUCTION_CUT_DATE'
AND PA.ATTRB_AN_VALUE >=cast(TRUNC(SYSDATE-1)+ INTERVAL '00:00:00' HOUR TO SECOND AS timestamp)
AND pa.ATTRB_AN_VALUE < CAST(TRUNC(SYSDATE)+ INTERVAL '00:00:00' HOUR TO SECOND AS timestamp)
Sample value for pa.attrb_an_value : 03-FEB-21 23:43:26,000000
But I get the following error.
hour must be between 1 and 12
you can first convert the date into timestamp. Instead of ATTRB_AN_VALUE please use
to_timestamp(substr(ATTRB_AN_VALUE,1,18),'DD.MM.YYYY HH24:MI:SSFF3')
This will convert the value into 03-FEB-21 11.43.26.000000 PM and it will eliminate the error.
Since the column attrb_an_value is not a DATE or TIMESTAMP but a VARCHAR2, you cannot compare it to a date without some casting. The TO_TIMESTAMP function will take a string and convert that to a timestamp value with a given format mask.
SELECT
*
FROM
piece p,
piece_attrb pa
WHERE
p.piece_num_id = pa.piece_num_id AND
pa.attrb_code = 'PRODUCTION_CUT_DATE' AND
TO_TIMESTAMP(pa.attrb_an_value,'DD-MON-YY HH24:MI:SS,FF6') >= TRUNC(systimestamp,'DD') - INTERVAL '1' DAY AND
TO_TIMESTAMP(pa.attrb_an_value,'DD-MON-YY HH24:MI:SS,FF6') < TRUNC(systimestamp,'DD')
Note 1: This will fail as soon as a row does not contain a string matching the DD-MON-YY HH24:MI:SS,FF6 format mask.
Note 2: As others pointed out, this is a serious design flaw. No date or timestamp data should be stored in VARCHAR2 columns.
I think your problem is about formatting the date. Here's the correct formatting. Also, I thought that you wanted result set that contained PA's ATTRB_AN_VALUE values in between the beginning of yesterday and today. So, the answer contains the simplified version of compared dates.
SELECT * FROM PIECE P, PIECE_ATTRB PA WHERE P.PIECE_NUM_ID=PA.PIECE_NUM_ID
AND PA.ATTRB_CODE='PRODUCTION_CUT_DATE'
AND to_timestamp(PA.ATTRB_AN_VALUE,'DD-MON-RR HH24:MI:SS,FF') >=to_timestamp(trunc(sysdate-1))
AND to_timestamp(pa.ATTRB_AN_VALUE,'DD-MON-RR HH24:MI:SS,FF') < to_timestamp(trunc(sysdate));
NOTE: I am running my query in Qubole's presto and sql command engine.
I am trying to convert my string timestamp to just date but none of the options are working out.
My string timestamp looks like 2017-03-29 10:32:28.0
and I want to have it like 2017-03-29
I have tried following queries to convert this string timestamp to retrieve date
1. select cast(created as date) from table1
Value cannot be cast to date: 2017-05-26 17:23:58.0
2. select cast(from_iso8601_timestamp(created) as date) from table1
Invalid format: "2014-12-19 06:06:36.0" is malformed at " 06:06:36.0"
3. select date(created) from table1
Value cannot be cast to date: 2012-10-24 13:50:00.0
How I can convert this timestamp to date in presto/sql?
As far as explained in the documentation, prestoDB seems to expect timestamps in a format '2001-08-22 03:04:05.321', and dates in a '2001-08-22'.
One solution would be to use a string function to extract the relevant part of the string before converting it. We know that the date part is located before the first space in the string, so.
If you need the date part as a string datatype:
split_part(created, ' ', 1)
If you need the date part as a date datatype:
cast(split_part(created, ' ', 1) as date)
You can try to use one of the following solutions:
SELECT
'2017-03-29 10:32:28.0' AS input_string,
DATE(date_parse('2017-03-29 10:32:28.0', '%Y-%m%-%d %H:%i:%s.%f')) AS solution_1,
DATE(try_cast('2017-03-29 10:32:28.0' as timestamp)) AS solution_2
I have a little query that strips the date from the datetime field but when I try to convert it from GMT to CST it readds the date. Is there a better way to do this?
Location table:
arrival
4-6-2018 12:35:43
SELECT arrival
FROM(
SELECT CONVERT(VARCHAR(8), arrival))
FROM locations
)a
This query will give me this result:
12:35:43
SELECT (DATEADD(hour,-5,arrival))
FROM(
SELECT CONVERT(VARCHAR(8), arrival))
FROM locations
)a
4-6-2018 12:35:43
This query will give readd the date. How can I remove the date and then do the dateadd function without it readding the date
arrival seems to be a DateTime, which always carries a date part. You need a time instead, supported by SQL Server 2008+:
cast(DATEADD(hour,-5,arrival) as time)
To quote from DATEADD (Transact-SQL) - Return Types:
The return data type is the data type of the date argument, except for string literals. The return data type for a string literal is datetime. An error will be raised if the string literal seconds scale is more than three positions (.nnn) or contains the time zone offset part.
Emphasis my own.
As you are passing a string (varchar), then DATEADD is returning a datetime.
Like those in the comments have said, if you use the correct data type (time) this problem goes away:
SELECT DATEADD(HOUR, -5,CONVERT(time,Arrival))
FROM (VALUES('4-6-2018 12:35:43'),('4-6-2018 07:35:43'),('4-6-2018 03:35:43')) V(Arrival)
Probably this is what you are asking for:
SELECT Convert(Varchar(8), DATEADD(hour,-5,arrival), 108)
FROM locations;
Note: This is compatible with SQL server versions that doesn't have Time datatype too.
I am using Business Objects, which runs on top of an Oracle SQL database. I do not have access to PL or any kind of SQL command line, and I do not have write access to the database. I can only run queries as single commands, requiring a defined set of columns to be output.
I am able to take data from user prompts, which appear in the SQL as:
#variable('Prompt1')
For example I can say:
SELECT
A.SomeDate
FROM
A
WHERE
A.SomeDate BETWEEN #variable('Start') AND #variable('End Date')
This is easy enough. It runs; requests the user to input some dates; and then returns all matches which are between them (inclusive).
The problem is, however, the users will be using Business Objects' "Infoview" system to run the queries, and the prompt system presents a date picker - which by default includes the time portion of the date ("01/01/2016 12:00:00 AM").
If the user does not delete the time portion, it can cause records to be missed, if the SomeDate value falls outside the selected time. For example, if I want to take all records of TODAY, then I technically want everything between 00:00:00 (midnight) and 23:59:59.
What I would really like to be able to do is use TRUNC around the query variable, as follows:
WHERE
A.SomeDate BETWEEN TRUNC(#variable('Start')) AND TRUNC(#variable('End Date'))
... however this causes a compilation error: "inconsistent datatypes: expected DATE got NUMBER". I don't know why Oracle would treat a prompt as a number datatype before it has compiled.
Does anyone know how I can take the #variable value and convert it into something that I will be able to truncate to a date value?
I'm therefore trying to figure out a way round this. One thing I had in mind was if I could possibly take the prompt variable and convert it explicitly into a date, using TO_DATE
Edit: it has been pointed out to me that TRUNC will have no effect, as the "12:00:00 AM" is already midnight. Therefore I think I have misunderstood TRUNC. It appears that it truncates it to midnight: whereas I thought it simply removed the time portion of the date altogether, meaning that matches would be returned at any time between 00:00:00 and 23:59:59.
What I really want is: if SomeDate has a time portion of, for example, 11:03 then how do I ensure that this will be included when an End Date prompt only specifies the day?
If you want to match SomeDate values between 00:00:00 on Start and 23:59:59 on End you can either adjust the end date to have that time instead of the default midnight, or use a range instead of between:
WHERE
A.SomeDate >= #variable('Start')
AND
A.SomeDate < #variable('End Date') + 1
The + 1 uses Oracle date arithmetic to give you the day after the variable value, so if the user picked "01/01/2016 12:00:00 AM" for both the start and end dates they would evaluate as 2016-01-01 00:00:00 and 2016-01-02 00:00:00 respectively. You can use the interval syntax if you prefer.
By using less-than for the upper limit you get all records where SomeDate is greater than or equal to the start date 2016-01-01 00:00:00 and less than the adjusted end date 2016-01-02 00:00:00 - which is the same as saying up to 2016-01-01 23:59:59. (Or if you has a timestamp column which has sub-second precision, up to 23:59:59.999...).
If the parser assumes the variable will be a string but it is actually a date - causing an 'inconsistent datatypes' error - then you could cast it to a date to satisfy the parser:
WHERE
A.SomeDate >= CAST(#variable('Start') AS DATE)
AND
A.SomeDate < CAST(#variable('End Date') AS DATE) + 1
or if it is actually passed as a string in the format you showed you can explicitly convert it:
WHERE
A.SomeDate >= TO_DATE(#variable('Start'), 'DD/MM/YYYY HH:MI:SS AM')
AND
A.SomeDate < TO_DATE(#variable('End Date'), 'DD/MM/YYYY HH:MI:SS AM') + 1
... making sure you have the correct format; from your example it could be DD/MM/YYYY or MM/DD/YYYY.
Try using TO_CHAR() and TO_DATE() together :
WHERE
A.SomeDate > TO_DATE(TO_CHAR(#variable('Prompt1'),'ddmmyyyy'),'ddmmyyyy')
First off, your problem is not coming from a time value in the prompt value, but rather the time value in SomeDate. Getting rid of that (making the date equal to midnight) will resolve the issue.
Your best bet, if you have the option of modifying the universe, is to create another object. I'm assuming you have an object named SomeDate whose SQL is a.somedate. Create another object, let's call it SomeDateOnly with a definition of trunc(a.somedate)* **.
Since SomeDateOnly will always be a midnight value, you can use Equal To with your prompts, which will produce SQL like:
trunc(a.somedate) = #variable('Prompt1')
which, when rendered by WebI, will produce:
trunc(a.somedate) = '16-08-2016 00:00:00'
This will return all records with a.somedate between 8/16/2016 at 00:00:00 and 8/16/2016 23:59:59.
Of course, you can use BETWEEN to select a range of dates:
trunc(a.somedate) between #variable('Start Date') and #variable('End Date')
Even if you don't have access to the universe, you can still use the above syntax by modifying WebI's generated SQL. (I'm assuming that's what you've been doing, anyway).
If the above works for you, then the following is irrelevant, but I wanted to address it anyway:
The reason for the "invalid number" error you were receiving is because of the way WebI formats dates for SQL. If you have this string in your query:
A.SomeDate = TRUNC(#variable('Prompt1'))
then WebI will replace the #variable(...) with a date string, and render it as the following before sending it to Oracle:
A.SomeDate = TRUNC('16-08-2016 00:00:00')
This, of course, makes no sense to the TRUNC() function as there's nothing to tell it that it's actually a date value.
You could to_date the prompt first, but you have to use the correct date format. WebI sets the nls_date_format for each session to a non-default format, so you would have to use:
A.SomeDate = TRUNC(to_date(#variable('Prompt1')),'dd-mm-yyyy hh24:mi:ss')
But again, this is irrelevant since you need to trunc somedate, not the prompt response values.
*Better still, rename SomeDate to SomeDateTime, and name the new object SomeDate
**This is pretty common - having multiple objects for the same source field. Sometimes you want the date/time value (for listing specific transactions), but sometimes you just need the date (for counting transactions by date). So having both available is very useful.
I'm working on a data warehouse project and would like to know how to (preferably in a Derived Column component in a Data flow) strip the date piece off of a SQL datetime record.
Once I have the datetime converted to just a time I am going to do a lookup on the time to find the related time record in a time dimension table.
Can someone give me a simple function to accomplish this inside a derived column transform?
Example: Transform a datetime such as "12/02/2008 11:32:21 AM" into simply "11:32:21 AM".
I would just do a cast to DT_DBTIME type (using Derived Column transform, or Convert type transform). DT_DBTIME contains just (hours, minutes, seconds) part of the date/time, so you'll get rid of the date part.
If you need to do this in a variable expression Michael's solution won't work, but you can use the following expression:
(DT_DATE)(DT_DBDATE)GETDATE()
(DT_DBDATE) converts the current date and time to a date only. But the new datatype is not compatiple with SSIS's datetime. Therefore you'll have to use (DT_DATE) for converting to a compatible type.
Courtesy of this solution belongs to Russel Loski who has posted it in his blog:
http://www.bidn.com/blogs/RussLoski/ssas/1458/converting-datetime-to-date-in-ssis
Actually if you reverse the first 2 expressions like this: (DT_DBDATE)(DT_DATE)GETDATE()
instead of (DT_DATE)(DT_DBDATE)GETDATE(), then you will TRUNCATE the time off the date field.
If the DT_DATE expression is before the DT_DBDATE expression, you will still have the time portion in your output, but it will be reset to all zeroes.
Ran into this with writing a report for a scheduling app, needed the time that was stored as part of a datetime data type. I formated the datetime as 0 which gives you this mon dd yyyy hh:miAM (or PM), and just did a substring of that which returned the time only in an AM/PM format.
Example below.
DECLARE #S DATETIME = GETDATE()
SELECT SUBSTRING(CONVERT(NVARCHAR(30), #S , 0) , 13 , 10) AS ApptTime
, CONVERT(NVARCHAR(30), #S , 0) AS ApptDate
I personally use a series of functions for this. E.g.:
ALTER FUNCTION [dbo].[TIMEVALUE]
(
#Datetime datetime
)
RETURNS datetime
AS
BEGIN
RETURN (#Datetime - CAST(ROUND(CAST(#Datetime AS float), 0, 1) AS datetime))
END
I'd love to claim all the credit but it should really go to this guy.