I have a table with a varchar(25) column that holds a date value. A typical value is '11/04/2017'.
This query returns 0 rows
select *
from myTable
where isdate(inputDate) = 0
I am trying to find a max on this, using a date sort.
This query returns the expected result
;with gooddates as
(
select
medcomfolder, PatientId, PatientBirthday, InputDate
from
myTable
where
isdate(inputDate) = 1
)
select max(convert(datetime, inputDate))
from gooddates
This query returns an error.
;with gooddates as
(
select
medcomfolder, PatientId, PatientBirthday, InputDate
from
dwhFuData
where
isdate(inputdate) = 1
)
select max(convert(date, inputdate))
from gooddates
This is the returned error
Msg 241, Level 16, State 1, Line 274
Conversion failed when converting date and/or time from character string
The difference between the 2 queries is that the first is converting to a dateTime while the latter is converting to a date.
At this point, I can move forward w/ the dateTime option, but I am left wondering what I am missing.
I have checked that there are no embedded spaces, and all the columns have a len(InputDate) = 10 (there is NO time data included)
I selected distinct values,put them in excel, and did a date function on each row. I was hoping to get a #VALUE on 1 row. All the rows worked.
So there is nothing silly like '02/31/2019' going on.
How can a dateTime conversion pass when a simple date conversion does not?
My guess is that you have values that include a time stamp following the date (based on the fact that isdate() is always zero).
If so, one simple solution would be to use convert(date, left(inputdate, 10)). Another solution uses try_convert():
try_convert(date, inputdate)
To find the offending values:
select inputdate
from dwhFuData
where try_convert(date, inputdate) is null and inputdate is not null;
Related
I am using SQL Server 2017. I am trying to handle strings from a free format field and either convert them to a date in the format of "dd/mm/yyyy" or if they are not in this format then simply display the text verbatim.
I need this in a VIEW so can not use SET LANGUAGE. Sounds simple using Convert and IsDate but does not seem to work.
So for the snippet of code below (remember this will be in a view), I want to read the text and if the string converts to a date (ie. is in the format dd/mm/yyyy then run the convert to a date as I need it in date format for Excel to pick up (via Connect SQL Server database)), and if it does not convert to a date then display the text as it is.
create table dateTest1
(
idx int,
dateStringTest varchar(15)
);
insert into dateTest1 (idx, dateStringTest)
values (1, '13/01/2021'), (2, 'no');
select
case
when isdate(convert(datetime, dateStringTest, 103)) = 1
then convert(datetime, dateStringTest, 103)
else dateStringTest
end as dtres
from
dateTest1
--where idx = 1
Error:
Msg 241, Level 16, State 1, Line 15
Conversion failed when converting date and/or time from character string.
This error happens for idx = 2. Idx = 1 works ok.
Any assistance with this would be greatly appreciated. Thanks in advance.
You need to cast your resulting date to a varchar. A case expression can only return a single data type and the order-of-precedence means it is still trying to convert the varchar values to datetime
select
case when isdate( dateStringTest) = 1
then Cast(convert(datetime, dateStringTest, 103) as varchar(10))
else dateStringTest
end as dtres
from dateTest1
You can compact into a single statement (the same order of precedence applies)
select IsNull(Cast(Convert(datetime,Try_Cast(dateStringTest as date),103) as varchar(10)),dateStringTest)
from datetest1
Based on the syntax you are using, I assume you are using SQL Server. You should add the appropriate tag to your question.
The only way to do that in one column is to leave it as text.
Your use of ISDATE() is incorrect. If dateStringTest is not a valid date, CONVERT() on SQL Server will throw an error like:
The conversion of a varchar data type to a datetime data type resulted in an out-of-range value.
That's why you get an error for idx=2
You say you "need it in date format for Excel to pick up". How Excel interprets it will probably depend on your locale settings. Assuming mm/dd/yyyy is valid in your locale, Excel probably already sees it as a date. But I assume that's not happening, so mm/dd/yyyy is not valid for your locale. For me, if I have this table in SQL...
create table dateTest1 (
id int identity(1,1) not null,
dateStringTest varchar(20)
)
insert dateTest1
values
('2020-01-01')
, ('2020-21-01')
, ('2020-01-21')
, ('1/1/2020')
, ('1/21/2020')
, ('21/1/2020')
, ('21/01/2020')
, ('other stuff')
...and query it from Excel, adding columns with the functions DATEVALUE, DAY, WEEKDAY, and YEAR, I get...
id
dateStringTest
DATEVALUE
DAY
WEEKDAY
YEAR
1
2020-01-01
43831
1
4
2020
2
2020-21-01
#VALUE!
#VALUE!
#VALUE!
#VALUE!
3
2020-01-21
43851
21
3
2020
4
1/1/2020
43831
1
4
2020
5
1/21/2020
43851
21
3
2020
6
21/1/2020
#VALUE!
#VALUE!
#VALUE!
#VALUE!
7
21/01/2020
#VALUE!
#VALUE!
#VALUE!
#VALUE!
8
other stuff
#VALUE!
#VALUE!
#VALUE!
#VALUE!
That seems to indicate that Excel is recognizing some of the values as dates.
What you need to do is try to convert the value to datetime, then if it fails report the original value. In both cases, output a string.
You should review the documentation for CONVERT().
Try this:
select dateStringTest
, coalesce(cast(try_convert(datetime, dateStringTest, 103) as varchar(20)), dateStringTest) as dtres1
, coalesce(convert(varchar(20), try_convert(datetime, dateStringTest, 103), 103), dateStringTest) as dtres2
, coalesce(convert(varchar(20), try_convert(datetime, dateStringTest, 103), 101), dateStringTest) as dtres3
from dateTest1
UPDATE
If you are considering only values matching the pattern dd/mm/yyyy as dates (so 17/01/2021 is a date and 17/1/2021 is not), this brute force method will work with SQL Server 2008 (compatibility level 100):
(Notice I updated my input, also, above.)
;
with a as (
select id
, dateStringTest
, SUBSTRING(dateStringTest, 7, 4) + '-' + SUBSTRING(dateStringTest, 4, 2) + '-' + SUBSTRING(dateStringTest, 1, 2) as converteDate
from dateTest1
)
select id
, cast(convert(datetime, dateStringTest, 103) as varchar(20)) as dtres
from a
where isdate(converteDate) = 1
union
select id
, dateStringTest as dtres
from a
where isdate(converteDate) = 0
I'm reporting out of a database that is using decimal(17,6) as the datatype for a date field. For example, the current date/time in this field would be 20210820.171900. Unusual, but whatever. I need to convert the original date field from decimal(17,6) to datetime. This is what I have:
SELECT convert(datetime, convert(varchar,convert(int, lastmoddatetime)), 0)
from Table1
The above statement works correctly as long as none of the records have a value of zero in this column. Unfortunately, the column value defaults to zero (0.000000) if no date has been calculated for it. Whenever a column has a zero value, I get the following error:
Conversion failed when converting date from character string.
How can I overcome this issue? Ultimately, I'm needing to apply a dateadd function to the lastmoddatetime field.
Note: Before you suggest changing the column definition, this database originated in the 1990's and I'm not allowed to make any changes to the database structure.
You can use NULLIF to null out those values
convert(datetime, convert(varchar(15), convert(int, NULLIF(lastmoddatetime, 0.0))), 0)
Either use TRY_CONVERT or CASE - depending how you want to handle the zero case.
SELECT
-- If desiring null for 0 and SQL Server 2012+
TRY_CONVERT(date, CONVERT(varchar, CONVERT(int, lastmoddatetime)), 0)
, CASE WHEN lastmoddatetime <> 0
-- If desiring some other valid date or < SQL Server 2012
THEN CONVERT(date, CONVERT(varchar, CONVERT(int, lastmoddatetime)), 0)
ELSE NULL /* Whatever valid datetime value you want */ END
FROM (
VALUES (20210820.171900), (0.0)
) x (lastmoddatetime);
I note that this ignores the time component - so am converting to a date not datetime above. If you need to handle the time component you need to update your question.
Yet another option.
You can thin it out a bit by using left() and try_convert()
Example
Declare #YourTable table (lastmoddatetime numeric(17,6))
Insert into #YourTable values
(20210820.171900)
,(0.0)
Select AsDate = try_convert(date,left(lastmoddatetime,8))
,AsDateTime = try_convert(datetime,left(lastmoddatetime,8))
From #YourTable
Results
AsDate AsDateTime
2021-08-20 2021-08-20 00:00:00.000
NULL NULL
use
convert(datetime,convert(int,lastmoddatetime),0)
I have a varchar column in my database table, on the row I would like to return it is populated as '2018-12-26T00:00:00.000' (quotes mine, not included in actual value). When I try to query for this value whenever it is a valid date, e.g.
SELECT
myValue
FROM
myTable
WHERE
ISDATE(myValue) = 1
it returns properly. However, I need this value to be converted to DATE. When I try something like this:
SELECT
CAST(myValue AS DATE) AS myValueFormatted
FROM
myTable
WHERE
ISDATE(myValue) = 1
I get an error
Conversion failed when converting date and/or time from character string
Is there any other way I can convert this varchar value to Date?
UPDATE: I've noticed through trying some different things, the query seems to be fine with me using the value as a date for anything (DATEDIFF, CONVERT back to string, etc.) in the select portion, but trying to do anything with it in the WHERE clause causes the error. To ensure nothing else is interfering, I created a temp table with only 1 row with the data value above, and running the query just against that one value gives the error
UPDATE 2: Ok, I have no idea why this fixes it, but this is what I found. When I run
SELECT
myValue
FROM
myTable
WHERE
TRY_CONVERT(DATE, myValue) IS NOT NULL
it returns EXACTLY the same values as
SELECT
myValue
FROM
myTable
WHERE
ISDATE(myValue) = 1
However, when I then add AND CAST(myValue AS DATE) < GETDATE() to each WHERE clause, only the first one works. I understand why TRY_CONVERT is safer to use, I'm still not sure why it works over GETDATE()
I can't reproduce your error...
declare #dt varchar(256) = '2018-12-26T00:00:00.000'
select cast(#dt as date)
So, there must be another rogue value in there that can't be converted.
To identify what value is causing the issue on versions < 2012, run this:
SELECT
myValue
FROM myTable
WHERE
ISDATE(myValue) = 0
Note, ISDATE is deterministic only if you use it with the CONVERT function, if the CONVERT style parameter is specified, and style is not equal to 0, 100, 9, or 109.
For 2012 onward, use TRY_CONVERT
SELECT
*
FROM myTable
WHERE
TRY_CONVERT(date, myValue) IS NULL
You could also just try something like this:
SELECT CAST(LEFT(MyValue, 10) AS DATE)
If it still doesn't work, you have some formatting issues with your data.
This helped me....
CAST string as varchar(30) then cast the varchar as datetime2
CAST(CAST(REPLACE(['Timestamp' ],'''','') AS varchar(30)) as datetime2)
I have a query I am running in redshift that produces an error when I try to compare two dates. I have determined this is due to a data problem where the dates are VARCHAR and some are empty strings. The best solution is clearly to fix this at the source, but while trying to build a work around, I stumbled upon some very odd behavior.
To get around, I preselect the dates that are not empty strings, and cast as dates, then convert to integer date format (YYYYMMDD) and convert to INT. This runs fine. However, if I try to compare this with an integer in a WHERE clause, the query crashes with a data type error.
Here is a toy version of the working query
SELECT
date_id,
COUNT(*)
FROM
(
SELECT
CONVERT(int, date_id) AS date_id
FROM
(
SELECT
DATE_PART('year', start_dttm)*10000+DATE_PART('month', start_dttm)*10+DATE_PART('day', start_dttm) AS date_id
FROM
(
SELECT
CAST(start_dttm AS DATETIME) AS start_dttm
FROM
sfe.calendar_detail
WHERE
start_dttm <> ''
) cda
) cdb
) cd
GROUP BY
date_id
;
And here is the failed query
SELECT
date_id,
COUNT(*)
FROM
(
SELECT
CONVERT(int, date_id) AS date_id
FROM
(
SELECT
DATE_PART('year', start_dttm)*10000+DATE_PART('month', start_dttm)*10+DATE_PART('day', start_dttm) AS date_id
FROM
(
SELECT
CAST(start_dttm AS DATETIME) AS start_dttm
FROM
sfe.calendar_detail
WHERE
start_dttm <> ''
) cda
) cdb
) cd
WHERE
date_id >= 20170920
GROUP BY
date_id
;
As I mentioned above, the correct solution is to fix the data type and count empty dates as Nulls not empty strings, but I am very curious as to why the second query crashes on an invalid data type error.
Many Thanks!
Edit:
Here is the error
ERROR: Invalid digit, Value '1', Pos 0, Type: Integer
DETAIL:
-----------------------------------------------
error: Invalid digit, Value '1', Pos 0, Type: Integer
code: 1207
context:
query: 2006739
location: :0
process: query0_39 [pid=0]
-----------------------------------------------
Rather than converting dates to the human-readable YYYYMMDD format, it is always better to keep them as DATE or TIMESTAMP format. This way, date operations can be easily performed (eg adding 5 days to a date). You can still do easy comparison operators by using 'YYYYMMDD'::DATE.
Given that you are converting from a String, and casting to a Date seems to work, and that you have some empty strings, use this to convert it to a date:
SELECT
NULLIF(start_dttm, '')::DATE AS dt
FROM sfe.calendar_detail
WHERE dt > '20170920'::DATE
This will return a NULL if the string is empty, and a Date if it contains a date that could be converted.
I have a column which contains data but the problem is that this column has data type of varchar(50) and it has to be this due to some reasons,now what i want to do is while selecting data from table , i want to treat this column as date so that i can use it in where clause. i am using the code below for converting it yo date , but it converts some values and then gives an error
this is my sample data
8/1/2002
6/9/2001
14/9/2001
26/7/2001
14/12/2001
21/1/2002
29/4/2001
7/5/2001
9/11/2001
16/7/2001
select CONVERT(date,sowingDate,103) from tblAgriculture_staging
I have tried which differnt version of date format e.g 103,105 etc
but still it converts some values but error comes on some values and query execution stops
Try this:
SET DATEFORMAT dmy;
select case when isdate(sowingDate) = 1 then CONVERT(date,sowingDate,103) end [date] from tblAgriculture_staging
or (if you are using sql 2012)
SET DATEFORMAT dmy;
select case when TRY_CONVERT(date, sowingDate) IS NOT NULL then CONVERT(date,sowingDate,103) end [date] from tblAgriculture_staging
but this solution hides (convert to NULL) all dates that are wrong. You can reverse the condition first and find/fix all rows with incorrect date (i.e. 31/02/2013) and then use this queries to show only valid dates
SQLFiddle
but it converts some values and then gives an error this is my sample
data
because some data are in invalid format or contains incorrect symbols.
Try this:
select CONVERT(date,ltrim(rtrim(sowingDate)), 103) from tblAgriculture_staging
or examine your values:
select ISDATE(sowingDate) as IsDate, sowingDate, CASE WHEN ISDATE(sowingDate)=1 THEN CONVERT(date,ltrim(rtrim(sowingDate)), 103) ELSE NULL END from tblAgriculture_staging
This is slightly crappy, but so is storing dates as varchar.
this is code that has worked for me in the past where i had some dates with 4 digit years and some with 2 digit years.
where (TRY_CONVERT(Datetime2,LTRIM(RTRIM([INVC DTE])),1)>=#From
AND TRY_CONVERT(Datetime2,LTRIM(RTRIM([INVC DTE])),1)<=#To)
OR (TRY_CONVERT(Datetime2,LTRIM(RTRIM([INVC DTE])),101)>=#From
AND TRY_CONVERT(Datetime2,LTRIM(RTRIM([INVC DTE])),101)<=#To)
SQL Server 2012 + Only
This assumes you have cleaned up anything that actually just isn't a date...
This will return all the dates that are not actually dates.
select sowingDate from tblAgriculture_staging where isdate(sowingDate)=0