March 20, 2013 at 8:00 am
All,
I'm having strange issue trying to do the following which I thought was just going to be straight forward
select top 5 machine.name, machine.model, machine.scantime
'INSERT INTO [SMS_000].[dbo].[BMCMachines] ([ComputerName],[MachineModel],[stime]) VALUES (' + '''' + machine.name + ''',' + '''' + machine.model + ''',' + convert(datetime,machine.scantime,112) + ''')' from machine
when I run the query comes back with following error
Msg 241, Level 16, State 1, Line 1
Conversion failed when converting date and/or time from character string.
I can confirm that machine.scantime is "DateTime" column.
March 20, 2013 at 8:46 am
denis.gendera (3/20/2013)
All,I'm having strange issue trying to do the following which I thought was just going to be straight forward
select top 5 machine.name, machine.model, machine.scantime
'INSERT INTO [SMS_000].[dbo].[BMCMachines] ([ComputerName],[MachineModel],[stime]) VALUES (' + '''' + machine.name + ''',' + '''' + machine.model + ''',' + convert(datetime,machine.scantime,112) + ''')' from machine
when I run the query comes back with following error
Msg 241, Level 16, State 1, Line 1
Conversion failed when converting date and/or time from character string.
I can confirm that machine.scantime is "DateTime" column.
Take a closer look at your Convert function and be sure that is what you are actually wanting to do.
March 25, 2013 at 10:08 pm
Just confirm if scantime is date or date time ???
also if scantime is already date, why convert it again..
March 26, 2013 at 4:05 am
denis.gendera (3/20/2013)
All,I'm having strange issue trying to do the following which I thought was just going to be straight forward
select top 5 machine.name, machine.model, machine.scantime
'INSERT INTO [SMS_000].[dbo].[BMCMachines] ([ComputerName],[MachineModel],[stime]) VALUES (' + '''' + machine.name + ''',' + '''' + machine.model + ''',' + convert(datetime,machine.scantime,112) + ''')' from machine
when I run the query comes back with following error
Msg 241, Level 16, State 1, Line 1
Conversion failed when converting date and/or time from character string.
I can confirm that machine.scantime is "DateTime" column.
The issue with the convert is that the 3rd arguement is the style and is used to set the style when you convert from a datetime to a varchar, not the other way around.
What is the datatype of the [BMCMachines].[stime] column because the other issue is that the convert style of 112 is for an ISO date with no time element to it??
March 26, 2013 at 6:30 am
please provide table structure
_______________________________________________________________
To get quick answer follow this link:
http://www.sqlservercentral.com/articles/Best+Practices/61537/
March 26, 2013 at 6:52 am
Steve JP (3/26/2013)
...The issue with the convert is that the 3rd arguement is the style and is used to set the style when you convert from a datetime to a varchar, not the other way around....
It works both ways as documented in BOL under CAST and CONVERT (Transact-SQL).
SELECT DateAsCharacter = CONVERT(VARCHAR(25),GETDATE(),100)
-- 'Mar 26 2013 12:49PM'
SELECT DateAsDate = CONVERT(DATETIME,'Mar 26 2013 12:49PM',100)
-- 2013-03-26 12:49:00.000
For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
March 26, 2013 at 6:58 am
select top 5 machine.name, machine.model, machine.scantime
'INSERT INTO [SMS_000].[dbo].[BMCMachines] ([ComputerName],[MachineModel],[stime]) VALUES (' + '''' + machine.name + ''',' + '''' + machine.model + ''',' + convert(datetime,machine.scantime,112) + ''')' from machine
This query is syntactically incorrect. Try putting a comma after machine.scantime. What is it that you are really trying to do? How about this?
INSERT INTO [SMS_000].[dbo].[BMCMachines]
([ComputerName],[MachineModel],[stime])
SELECT top 5
machine.name,
machine.model,
machine.scantime
FROM machine
For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
March 27, 2013 at 4:43 am
SELECT DateAsCharacter = CONVERT(VARCHAR(25),GETDATE(),100)-- 'Mar 26 2013 12:49PM'
SELECT DateAsDate = CONVERT(DATETIME,'Mar 26 2013 12:49PM',100)
-- 2013-03-26 12:49:00.000
Dont totally agree with this as the following will error which is as I expect because the style is not vailid
-- Just to 100% prove the style is invalid
SELECT DateAsCharacter = CONVERT(VARCHAR(25),GETDATE(),25510)
Msg 281, Level 16, State 1, Line 1
25510 is not a valid style number when converting from datetime to a character string.
But the following seems to want to ignore the style arguement.
SELECT[DateAsDate AM] = CONVERT(DATETIME,'Mar 26 2013 02:49AM' )
SELECT[DateAsDate PM] = CONVERT(DATETIME,'Mar 26 2013 02:49PM' )
SELECT[DateAsDate mm-dd-yy] = CONVERT(DATETIME,'Mar 26 2013 02:49PM' , 110 )
SELECT[DateAsDate yyyymmdd] = CONVERT(DATETIME,'Mar 26 2013 02:49PM' , 112 )
SELECT[DateAsDate No Such format] = CONVERT(DATETIME,'Mar 26 2013 02:49PM' ,25510) -- invalid style so Error? or Ignore?
DateAsDate AM = 2013-03-26 02:49:00.000
DateAsDate PM = 2013-03-26 14:49:00.000
DateAsDate mm-dd-yy = 2013-03-26 14:49:00.000
DateAsDate yyyymmdd = 2013-03-26 14:49:00.000
DateAsDate No Such format = 2013-03-26 14:49:00.000
March 27, 2013 at 5:06 am
Steve JP (3/27/2013)
SELECT DateAsCharacter = CONVERT(VARCHAR(25),GETDATE(),100)-- 'Mar 26 2013 12:49PM'
SELECT DateAsDate = CONVERT(DATETIME,'Mar 26 2013 12:49PM',100)
-- 2013-03-26 12:49:00.000
Dont totally agree with this as the following will error which is as I expect because the style is not vailid
-- Just to 100% prove the style is invalidSELECT DateAsCharacter = CONVERT(VARCHAR(25),GETDATE(),25510)
Msg 281, Level 16, State 1, Line 1
25510 is not a valid style number when converting from datetime to a character string.
- and your point is?
But the following seems to want to ignore the style arguement.
SELECT[DateAsDate AM] = CONVERT(DATETIME,'Mar 26 2013 02:49AM' )
SELECT[DateAsDate PM] = CONVERT(DATETIME,'Mar 26 2013 02:49PM' )
SELECT[DateAsDate mm-dd-yy] = CONVERT(DATETIME,'Mar 26 2013 02:49PM' , 110 )
SELECT[DateAsDate yyyymmdd] = CONVERT(DATETIME,'Mar 26 2013 02:49PM' , 112 )
SELECT[DateAsDate No Such format] = CONVERT(DATETIME,'Mar 26 2013 02:49PM' ,25510) -- invalid style so Error? or Ignore?
DateAsDate AM = 2013-03-26 02:49:00.000
DateAsDate PM = 2013-03-26 14:49:00.000
DateAsDate mm-dd-yy = 2013-03-26 14:49:00.000
DateAsDate yyyymmdd = 2013-03-26 14:49:00.000
DateAsDate No Such format = 2013-03-26 14:49:00.000
The style argument for converting a date as string in the format 'Mar 26 2013 12:49PM' to a datetime data type using CONVERT() is 100. Other style arguments are likely to fail. I'm sorry but I still don't see your point. None of this appears to prove anything either way. Have a quick read of CAST and CONVERT in BOL, specifically CONVERT relating to dates. The explanation is quite good. If you're still unsure, post back.
For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
March 27, 2013 at 6:30 am
Steve JP (3/26/2013)
....
The issue with the convert is that the 3rd arguement is the style and is used to set the style when you convert from a datetime to a varchar, not the other way around.
...
ChrisM@Work (3/27/2013)
....
The style argument for converting a date as string in the format 'Mar 26 2013 12:49PM' to a datetime data type using CONVERT() is 100. Other style arguments are likely to fail. I'm sorry but I still don't see your point. None of this appears to prove anything either way. Have a quick read of CAST and CONVERT in BOL, specifically CONVERT relating to dates. The explanation is quite good. If you're still unsure, post back......
Chris, I thought my example was clear in that the style code ( of 25510 ) that I first used doesnt exist. I know this & you know this, but others may not, as this is the newbies section hence I posted the code. The next code block shows that the Style arguement is often ignored for a conversion from varchar to datetime. It also doesnt have any affect on the time. ie the 103 (which is English) will take a datetime to a varchar as only a date of format dd/mm/yyyy. But from a varchar to datetime will take the time element if it exists as well so can be a bit misleading for some. How many times have people asked how to truncate the time element from a datetime value.
If the style arguement doest exist but cant be applied then it will error. Other times like I have shown it will ignore the style arguement if the actual data contradicts the style.
One of the most reliable methods I have found to enter datetime is yyyy-mm-dd hh:mm:ss:mmm (or the full iso style)
Another option is to use SET DATEFORMAT, although I'm not certain what lifespan this has long term.
I havent got all day for these forums and can only post while jobs are running etc. So yes some of the time my comments are a bit too brief, but mainly I try and post a bit more than "go read the BOLs".
March 27, 2013 at 7:42 am
Steve JP (3/27/2013)
Steve JP (3/26/2013)
....
The issue with the convert is that the 3rd arguement is the style and is used to set the style when you convert from a datetime to a varchar, not the other way around.
...
ChrisM@Work (3/27/2013)
....
The style argument for converting a date as string in the format 'Mar 26 2013 12:49PM' to a datetime data type using CONVERT() is 100. Other style arguments are likely to fail. I'm sorry but I still don't see your point. None of this appears to prove anything either way. Have a quick read of CAST and CONVERT in BOL, specifically CONVERT relating to dates. The explanation is quite good. If you're still unsure, post back......
Chris, I thought my example was clear in that the style code ( of 25510 ) that I first used doesnt exist. I know this & you know this, but others may not, as this is the newbies section hence I posted the code. The next code block shows that the Style arguement is often ignored for a conversion from varchar to datetime. It also doesnt have any affect on the time. ie the 103 (which is English) will take a datetime to a varchar as only a date of format dd/mm/yyyy. But from a varchar to datetime will take the time element if it exists as well so can be a bit misleading for some. How many times have people asked how to truncate the time element from a datetime value.
If the style arguement doest exist but cant be applied then it will error. Other times like I have shown it will ignore the style arguement if the actual data contradicts the style.
One of the most reliable methods I have found to enter datetime is yyyy-mm-dd hh:mm:ss:mmm (or the full iso style)
Another option is to use SET DATEFORMAT, although I'm not certain what lifespan this has long term.
I havent got all day for these forums and can only post while jobs are running etc. So yes some of the time my comments are a bit too brief, but mainly I try and post a bit more than "go read the BOLs".
Hi Steve
I guess I owe you an apology for the brevity of my last reply. Here's something I knocked up earlier this morning which should help - seeing something in action makes the concept far more likely to "stick" than reading about it.
Match a date string to a style argument for conversion to a DATETIME, and show the output when the reverse is performed, DATETIME to VARCHAR:
;WITH RawData ([Standard], DateString, Style) AS (
SELECT 'DEFAULT', 'Mar 27 2013 11:57AM',100 UNION ALL
SELECT 'U.S.', '03/27/2013',101 UNION ALL
SELECT 'ANSI', '2013.03.27',102 UNION ALL
SELECT 'British/French', '27/03/2013', 103 UNION ALL
SELECT 'German', '27.03.2013',104 UNION ALL
SELECT 'Italian', '27-03-2013',105 UNION ALL
SELECT '-', '27 Mar 2013',106 UNION ALL
SELECT '-', 'Mar 27, 2013',107 UNION ALL
SELECT 'Default + milliseconds', 'Mar 27 2013 11:57:23:607AM',109 UNION ALL
SELECT 'USA', '03-27-2013',110 UNION ALL
SELECT 'JAPAN', '2013/03/27',111 UNION ALL
SELECT 'ISO', '20130327',112 UNION ALL
SELECT 'Europe default + milliseconds', '27 Mar 2013 11:57:23:607',113
)
SELECT r.[Standard], x.*
FROM RawData r
CROSS APPLY (
SELECT
[String To DATETIME] = CONVERT(DATETIME,r.DateString,r.Style),
[DATETIME to String] = CONVERT(VARCHAR(26), GETDATE(),r.Style)
) x;
For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
March 27, 2013 at 8:26 am
Edit: Nevermind, should have kept reading before posting.
March 27, 2013 at 10:32 am
Tis why I avoid passing comment straight away 🙂
March 27, 2013 at 10:47 am
Chris,
Often the style gets ignored or overridden by the character string being in a different state. It also keeps the time element which seems to me slighty at odds compared to convert this datetime to a date string.
I will try and encourage develoeprs to use a datetime type rather than passing a string and hoping that the database will alter it to something correct. XML has a proper datetime type as well.
And yes I have had my fingers burnt by the odd load coming in with some dates being in one format and another being in a different format, in the same file.
Viewing 14 posts - 1 through 13 (of 13 total)
You must be logged in to reply to this topic. Login to reply