Cast as INT check SQL - sql

I need to update a column in a table. But only where the cast to an INT fails.
I have the following so far - but this updates all the records.
begin try
select cast(customerid as int) from Table_staging;
end try
begin catch
update Table_staging
set incorrectformat = 0
end catch

update Table_staging
set incorrectformat = 0
where not customerid like '%[^0-9]%'
should be sufficient. Basically, we mark incorrectformat for any row where customerid is not a string that contains any number of characters, then a character not in the set 0-9, than any number of characters.
I.e. the values that don't match this are precisely the ones only containing digits.
And the main issue with ISNUMERIC is that it answers a question that I don't believe anyone would ever, rightfully, ask - "Can this character string be converted to any of the numeric data types? I don't care which of those types it can be converted to, and there's no need to tell me which ones in response either"

Related

SQL Server: Return a string in a specific format

In TSQL, I need to format a string in a predefined format.
For eg:
SNO
STRING
FORMAT
OUTPUT
1
A5233GFCOP
*XXXXX-XXXXX
*A5233-GFCOP
2
K92374
/X-000XXXXX
/K-00092374
3
H91543987
XXXXXXXXX
H91543987
I am trying with FORMATMESSAGE() built in function.
For ex:
FORMATMESSAGE('*%s-%s','A5233','GFCOP')
FORMATMESSAGE('/%s-000%s','K','92374')
FORMATMESSAGE('%s','H91543987')
I am able to get the first argument by replace function but issue is second/third/fourth/.. arguments.
I don't know how to count respective X's between the various delimiters, so that I can use substring to pass in second/third/.. arguments. If I can count the respective # of X's from the Format column, I feel using substring we can get it but not sure how to count the respective X's.
Please let me know how to get through it or if there is any other simple approach.
Appreciate your help.
Thanks!
It's in theory quite simple, could probably be done set-based using string_split however that's not ideal as the ordering is not guaranteed. As the strings are fairly short then a scalar function should suffice. I don't think it can use function in-lining.
The logic is very simple, create a counter for each string, loop 1 character at a time and pull a character from one or the other into the output depending on if the format string is an X or not.
create or alter function dbo.fnFormatString(#string varchar(20), #format varchar(20))
returns varchar(20)
as
begin
declare #scount int=1, #fcount int=1, #slen int=len(#string), #flen int=Len(#format), #output varchar(20)=''
while #scount<=#slen or #fcount<=#slen
begin
if Substring(#format,#fcount,1)='X'
begin
set #output+=Substring(#string,#scount,1)
select #scount+=1, #fcount +=1
end
else
begin
set #output+=Substring(#format,#fcount,1)
set #fcount +=1
end
end
return #output
end;
select *, dbo.fnFormatString(string, [format])
from t
See working Fiddle

SQL: Replacing dates contained within a text string

I am using SQL Server Management Studio 2012. I work with medical records and need to de-identify reports. The reports are structured in a table with columns Report_Date, Report_Subject, Report_Text, etc... The string I need to update is in report_text and there are ~700,000 records.
So if I have:
"patient had an EKG on 04/09/2012"
I need to replace that with:
"patient had an EKG on [DEIDENTIFIED]"
I tried
UPDATE table
SET Report_Text = REPLACE(Report_Text, '____/___/____', '[DEIDENTIFED]')
because I need to replace anything in there that looks like a date, and it runs but doesn't actually replace anything, because apparently I can't use the _ wildcard in this command.
Any recommendations on this? Advance thanks!
You can use PATINDEX to find the location of Date and then use SUBSTRING and REPLACE to replace the dates.
Since there may be multiple dates in the Text you have to run a while loop to replace all the dates.
Below sql will work for all dates in the form of MM/DD/YYYY
WHILE EXISTS( SELECT 1 FROM dbo.MyTable WHERE PATINDEX('%[0-9][0-9]/[0-9][0-9]/[0-9][0-9][0-9][0-9]%',Report_Text) > 0 )
BEGIN
UPDATE t
SET Report_Text = REPLACE(Report_Text, DateToBeReplaced, '[DEIDENTIFIED]')
FROM ( SELECT * ,
SUBSTRING(Report_Text,PATINDEX('%[0-9][0-9]/[0-9][0-9]/[0-9][0-9][0-9][0-9]%',Report_Text), 10) AS DateToBeReplaced
FROM dbo.MyTable AS a
WHERE PATINDEX('%[0-9][0-9]/[0-9][0-9]/[0-9][0-9][0-9][0-9]%',Report_Text) > 0
) AS t
END
I have tested the above sql on a dummy table with few rows.I don't know how it will scale for your data but recommend you to give it a try.
To keep it simple, assume that a number represents an identifying element in the string so look for the position of the first number in the string and the position of the last number in the string. Not sure if this will apply to your entire set of records but here is the code ...
I created two test strings ... the one you supplied and one with the date at the beginning of the string.
Declare #tstString varchar(100)
Set #tstString = 'patient had an EKG on 04/09/2012'
Set #tstString = '04/09/2012 EKG for patient'
Select #tstString
-- Calculate 1st Occurrence of a Number
,PATINDEX('%[0-9]%',#tstString)
-- Calculate last Occurrence of a Number
,LEN(#tstString) - PATINDEX('%[0-9]%',REVERSE(#tstString))
,CASE
-- No numbers in the string, return the string
WHEN PATINDEX('%[0-9]%',#tstString) = 0 THEN #tstString
-- Number is the first character to find the last position and remove front
WHEN PATINDEX('%[0-9]%',#tstString) = 1 THEN
CONCAT('[DEIDENTIFIED]',SUBSTRING(#tstString, LEN(#tstString)-PATINDEX('%[0-9]%',REVERSE(#tstString))+2,LEN(#tstString)))
-- Just select string up to the first number
ELSE CONCAT(SUBSTRING(#tstString,1,PATINDEX('%[0-9]%',#tstString)-1),'[DEIDENTIFIED]')
END AS 'newString'
As you can see, this is messy in SQL.
I would rather achieve this with a parser service and move the data with SSIS and call the service.

Replacing a calculated field column with a Blank when a NULL is returned?

I'm doing a simple query that uses the DateDiff function to find the number of days between the dates. However, with regards to certain instances, I'd like to populate a blank field (not a null).
Something like this is what I currently have, and it seems to work fine (but it populates a null).
[Test (Years)] = CASE WHEN TYPE IN ('A','B')
THEN NULL ELSE IsNull(CONVERT(decimal(28,12),
(DATEDIFF(d,#StartDate,ExpirationDate)))/365,0) END
Now if I try something like this... which tries to convert all TYPE A and B to populate a blank, I'll get the following error message: Error converting data type varchar to numeric.
[Test (Years)] = CASE WHEN TYPE IN ('A','B')
THEN '' ELSE IsNull(CONVERT(decimal(28,12),
(DATEDIFF(d,#StartDate,ExpirationDate)))/365,0) END
Is there a simple thing I'm missing? I've tried doing the calcualtions without converting to a decimal, but it doesn't seem to work. Any ideas? Thanks
CASE is an expression that returns exactly one value and all of the branches must yield compatible types. A string (even a blank string) is not compatible with a decimal, so you need to do something like:
CASE WHEN ... THEN '' ELSE
CONVERT(VARCHAR(32), COALESCE(CONVERT(DECIMAL(23,12), ... ,0)) END
Note that this hack will only work if you are presenting the data to an end user. If you are trying to store this data in a column or use it in other calculations, it too will be tripped up by the blank string. A number can't be a blank string:
DECLARE #i INT = '';
SELECT #i;
Result:
0
So, if you don't want "empty" numerics to be interpreted as 0, stop being afraid of NULL and if you are dealing with this at presentation time, have the presentation layer present a blank string instead of NULL.

In SQL Server 2008, how do I check if a varchar parameter can be converted to datatype money?

I've got a stored procedure I use to insert data from a csv. The data itself is a mix of types, some test, some dates, and some money fields. I need to guarantee that this data gets saved, even if it's formatted wrong, so, I'm saving them all to varchars. Later, once the data's been validated and checked off on, it will be moved to another table with proper datatypes.
When I do the insert into the first table, I'd like to do a check that sets a flag (bit column) on the row if it needs attention. For instance, if what should be a money number has letters in it, I need to flag that row and add the column name in an extra errormsg field I've got. I can then use that flag to find and highlight for the users in the interface the fields they need to edit.
The date parameters seem to be easy, I can just use IF ISDATE(#mydate) = '0' to test if that parameter could be converted from varchar to datetime. But, I can't seem to find an ISMONEY(), or anything that's remotely equivalent.
Does anyone know what to call to test if the contents of a varchar can legitimately be converted to money?
EDIT:
I haven't tested it yet, but what do you think of a function like this?:
CREATE FUNCTION CheckIsMoney
(
#chkCol varchar(512)
)
RETURNS bit
AS
BEGIN
-- Declare the return variable here
DECLARE #retVal bit
SET #chkCol = REPLACE(#chkCol, '$', '');
SET #chkCol = REPLACE(#chkCol, ',', '');
IF (ISNUMERIC(#chkCOl + 'e0') = '1')
SET #retVal = '1'
ELSE
SET #retVal = '0'
RETURN #retVal
END
GO
Update
Just finished testing the above code, and it works!
money is decimal in effect, so you test this way
Don't use ISNUMERIC out of the box though: it's unreliable. Use this:
ISNUMERIC(MyCOl + 'e0')
Note, if you have 6 decimal places then it will be lost on conversion to money
Other question with more info why: How to determine the field value which can not convert to (decimal, float,int) in SQL Server
Edit:
Can do it in one line if you want
ISNUMERIC(REPLACE(REPLACE(#chkCOl, '$', ''), ',', '') + 'e0')

Conditionally branching in SQL based on the type of a variable

I'm selecting a value out of a table that can either be an integer or a nvarchar. It's stored as nvarchar. I want to conditionally call a function that will convert this value if it is an integer (that is, if it can be converted into an integer), otherwise I want to select the nvarchar with no conversion.
This is hitting a SQL Server 2005 database.
select case
when T.Value (is integer) then SomeConversionFunction(T.Value)
else T.Value
end as SomeAlias
from SomeTable T
Note that it is the "(is integer)" part that I'm having trouble with. Thanks in advance.
UPDATE
Check the comment on Ian's answer. It explains the why and the what a little better. Thanks to everyone for their thoughts.
select case
when ISNUMERIC(T.Value) then T.Value
else SomeConversionFunction(T.Value)
end as SomeAlias
Also, have you considered using the sql_variant data type?
The result set can only have one type associated with it for each column, you will get an error if the first row converts to an integer and there are strings that follow:
Msg 245, Level 16, State 1, Line 1
Conversion failed when converting the nvarchar value 'word' to data type int.
try this to see:
create table testing
(
strangevalue nvarchar(10)
)
insert into testing values (1)
insert into testing values ('word')
select * from testing
select
case
when ISNUMERIC(strangevalue)=1 THEN CONVERT(int,strangevalue)
ELSE strangevalue
END
FROM testing
best bet is to return two columns:
select
case
when ISNUMERIC(strangevalue)=1 THEN CONVERT(int,strangevalue)
ELSE NULL
END AS StrangvalueINT
,case
when ISNUMERIC(strangevalue)=1 THEN NULL
ELSE strangevalue
END AS StrangvalueString
FROM testing
or your application can test for numeric and do your special processing.
You can't have a column that is sometimes an integer and sometimes a string. Return the string and check it using int.TryParse() in the client code.
ISNUMERIC. However, this accepts +, - and decimals so more work is needed.
However, you can't have the columns as both datatypes in one go: you'll need 2 columns.
I'd suggest that you deal with this in your client or use an ISNUMERIC replacement
IsNumeric will get you part of the way there. You can then add some further code to check whether it is an integer
for example:
select top 10
case
when isnumeric(mycolumn) = 1 then
case
when convert(int, mycolumn) = mycolumn then
'integer'
else
'number but not an integer'
end
else
'not a number'
end
from mytable
To clarify some other answers, your SQL statement can't return different data types in one column (it looks like the other answers are saying you can't store different data types in one column - yours are all strign represenations).
Therefore, if you use ISNUMERIC or another function, the value will be cast as a string in the table that is returned anyway if there are other strigns being selected.
If you are selecting only one value then it could return a string or a number, however your front end code will need to be able to return the different data types.
Just to add to some of the other comments about not being able to return different data types in the same column... Database columns should know what datatype they are holding. If they don't then that should be a BIG red flag that you have a design problem somewhere, which almost guarantees future headaches (like this one).