SQL Server 2012 :: Premature Casting On 0 Rows Insert
Feb 28, 2015
I have encountered some weird behaviour. Code that has been working for "eternities" suddenly started to fail. I couldn't recreate it on any other machines.
I have written a sample script to illustrate the issue:
INSERT INTO #t_parcels (parcel_id, current_pos, end_pos)
SELECT1, 100000000000.0, 900000000000.0
[Code] ..
The last insert crashed with :"Arithmetic overflow error converting expression to data type int", even though there are no rows that satisfies the condition!
This is due to "diff" column having wrong datatype, BUT, the insert had no hits in the database. So how can inserting 0 rows crash with incorrect datatype?
I even copied the select so it was ran before the insert, in in that case, the SELECT completed successfully.
When i changed datatype in the table, the error went away, but I'm still curious what led to the error.
In a t-sql 2012 sql script, I have the following script, that only works for a few records since the value of TST.dbo.LockCombination.seq only contains the value or 1 in most cases. Basically for every join listed below, there should be 5 records where each record has a distinct seq value of 1, 2, 3, 4, and 5. Thus my goal is to determine how to add the missing rows to the TST.dbo.LockCombination where there are no rows for seq values of between 2 to 5. I would like to know how to insert the missing rows and then do the following update statement. Thus can you show me the sql on how to add the rows for at least one of the missing sequence numbers?
UDATE LKC SET LKC.combo = lockCombo2 FROM [LockerPopulation] A JOIN TST.dbo.School SCH ON A.schoolnumber = SCH.type JOIN TST.dbo.Locker LKR ON SCH.schoolID = LKR.schoolID AND A.lockerNumber = LKR.number JOIN TST.dbo.Lock LK ON LKR.lockID = LK.lockID JOIN TST.dbo.LockCombination LKC ON LK.lockID = LKC.lockID WHERE LKC.seq = 2
A normal select statement looks like the following:
select * from TST.dbo.Locker LKR JOIN TST.dbo.Lock LK ON LKR.lockID = LK.lockID JOIN TST.dbo.LockCombination LKC ON LK.lockID = LKC.lockID where LKR.number in (000,001,1237)
In case you need the ddl statements for the tables affected here are the ddl statements:
CREATE TABLE [dbo].[Locker]( [lockerID] [int] IDENTITY(1,1) NOT FOR REPLICATION NOT NULL, [schoolID] [int] NOT NULL, [number] [varchar](10) NOT NULL, [serialNumber] [varchar](20) NULL, [type] [varchar](3) NULL, [locationID] [int] NULL,
I am finding it difficult to find an example that allows for insertion of additional rows into a table, without dropping the table I'm inserting into. Or inserting specific values. Like this example..
[URL] ....
I have 6 table I am formatting the data to conform to the final table as I'm inserting it into, but none of these examples gives me the example needed. I am using SQL 2012.
<code> SELECT CONVERT(VARCHAR(50),[FName]) + ' ' + CONVERT(VARCHAR(50),[LName]) AS [CustName] ,CAST('ALARMCOM' as nvarchar(8)) as VendorName ,CONVERT(VARCHAR(25),[CUSTOMER_CS_ACCOUNT_NUMBER]) AS [Cust_ID] ,CONVERT(VARCHAR(40),[Charge_Description])as [ChargeType] ,CASE
I create a Trigger that allows to create news row on other table.
ALTER TRIGGER [dbo].[TI_Creation_Contact_dansSLX] ON [dbo].[_IMPORT_FILES_CONTACTS] AFTER INSERT AS
[code]...
But if I create an INSERT with 50 rows.. My table CONTACT and ADDRESS possess just one line.I try to create a Cursor.. but I had 50 lines with an AdressID and a ContactID differently, but an Account and an AccountId egual on my CONTACT table :
I know that if I have an nvarchar column I can use an equality like = N'supersqlstring' so it doesn't implicit cast as a varchar, like if I were to do ='supersqlstring'. And then I'll be a big SQL hero and all my stored procedures will run before a millisecond can whisper.
But if I'm comparing an nvarchar column to a varchar column, is it better to cast the varchar 'up' to an nvarchar or cast the nvarchar 'down' to a varchar?
For instance:
cast(a.varchar as nvarchar(100)) = an.nvarchar
or
cast(an.nvarchar as varchar(100)) = a.varchar
Leaving aside non-matching, like (at least I don't think) that SQL considers the varchar n to be equal to the nvarchar ń, what's the best way to handle this?
Pretend for a moment that each column contains a mixed letter and number ID with no accented or wiggly-squiggly Unicode characters; it's just designs clashing.
Is there a performance hitch doing it one way or another? Should I use COLLATE? Should one of the columns be altered?
I have created a trigger that is set off every time a new item has been added to TableA.The trigger then inserts 4 rows into TableB that contains two columns (item, task type).
Each row will have the same item, but with a different task type.ie.
I have a table with PO#,Days_to_travel, and Days_warehouse fields. I take the distinct Days_in_warehouse values in the table and insert them into a temp table. I want a script that will insert all of the values in the Days_in_warehouse field from the temp table into the Days_in_warehouse_batch row in table 1 by PO# duplicating the PO records until all of the POs have a record per distinct value.
Example:
Temp table: (Contains only one field with all distinct values in table 1)
Create table #table (id int identity , from_country varchar(20) , to_country varchar(20),noofdays int, datetravel datetime ) insert into #table(from_country,to_country,noofdays,datetravel) values ('Malaysia','India',2,getdate()-99), ('India','Singapore',4,getdate()-88), ('Singapore','China',5,getdate()-77), ('China','Japan',6,getdate()-66), ('Japan','USA',7,getdate()-55) select * from #table
I want to insert data to another table based on "noofdays" columns. If "noofdays" is 4 then 4 rows should inserted to new table with 1 day increment in "datetravel" column.
Ex : #table 1MalaysiaIndia22015-02-09 02:04:09.247 2IndiaSingapore42015-02-20 02:04:09.247
[code]...
In #table , 1st row noofdays is 2 , so in new table #table_new first 2 rows should inserted with 1 day increment in "datetravel" column.
There is a valuable script out there that will take the rows from a table and display INSERT STATEMENTS.
Good thing is this script converts the data to HEXADECIMAL ( or some other ) and we don't have to worry about dealing with apostrophies embedded in varchar fields.
Does anybody know how could I calculate the new date(adding @c to @bor subtracting @b to @c), having for example a declaration like this:DECLARE @a CHAR(12)DECLARE @b DATETIMEDECLARE @c INTSET @b='3.04.04';SET @c=6and to calculate the number of days between two dates with this kindof declaration(@a-@b or @b - @a):DECLARE @a CHAR(12)DECLARE @b DATETIMEDECLARE @c INTSET @a='12.2.04';SET @b='3.04.04';Thanks in advance.
I run the following statement and it will not update beyond 7 million plus rows and I have about 38 million to complete. I keep checking updated row counts and after 1/2 day it's still the same so I know something is wrong because it was rolling through no problem when I initiated it. I need to complete ASAP so it's adding to my frustration. The 'Acct_Num_CH' field is an encrypted field (fyi).
SET rowcount 10000 UPDATE [dbo].[CC_Info_T] SET [Acct_Num_CH] = 'ayIWt6C8sgimC6t61EJ9d8BB3+bfIZ8v' WHERE [Acct_Num_CH] IS NOT NULL WHILE @@ROWCOUNT > 0 BEGIN SET rowcount 10000 UPDATE [dbo].[CC_Info_T] SET [Acct_Num_CH] = 'ayIWt6C8sgimC6t61EJ9d8BB3+bfIZ8v' WHERE [Acct_Num_CH] IS NOT NULL END SET rowcount 0
I am writing a query to return some production data. Basically i need to insert either 1 or 2 rows into a Table variable based on a decision as to does the production part make 1 or 2 items ( The Raw data does not allow for this it comes from a look up in my database)
I can retrieve all the source data i need easily but when i come to insert it into the table variable i need to insert 1 record if its a single part or 2 records if its a twin part. I know could use a cursor but im sure there has to be an easier way !
Below is the code i have at the moment
declare @startdate as datetime declare @enddate as datetime declare @Line as Integer DECLARE @count INT
set @startdate = '2015-01-01' set @enddate = '2015-01-31'
Hi AllI want to insert rows from a table in a serverinto another table in another server usingINSERT SELECT command. For example :INSERT INTO Server1.database1.dbo.Tab1SELECT * FROM Server2.database2.dbo.Tab1WHERE Col1 = 1Can the command like this work ?If not, could you give me the solution ?Please help meThanks in advanceJohn Smile*** Sent via Developersdex http://www.developersdex.com ***Don't just participate in USENET...get rewarded for it!
I have a dataset that contains an EmployeeID, StartDate, EndDate, and Location. Each record tells me at which location and employee was during a payroll period (fortnightly). So the data looks like:
Employees can be at multiple locations during the two weeks. They can work at one location, stop working there, start working somewhere else, and then maybe go back to their old location. There are duplicate records here as each employee can be at the same location multiple times during the two week period. What I need to capture is the actual start and end date of an employee at each location for each 'assignment'. An assignment is defined as a continual period of employment at a location with gaps of no less than 4 days - if there is a gap of more than four days then that is classed as a new assignment.
Basically if the type code is 1 one then move the data to column phone1, if the type is 2 then move it to column phone2.
This would be fairly simple if we always have type codes 1 and 2. But sometimes we can have type 1 and not type 2, or we could have type 2 and not type1.
Right now we only have 2 type codes. But, in the future we could be adding a 3rd type. So that would add a 3rd column (phone3).
Below is my code that I have written. I move the data into a temp table then list it. I am thinking of making this a view to my table. It works just fine. My question is, is there a better and more efficient way of doing this?
CREATE TABLE #Contacts ( id INT PRIMARY KEY, phone1 VARCHAR(15), phone2 VARCHAR(15) )
-- Insert the records for type 1
INSERT INTO #Contacts SELECT id, phone_num, NULL FROM test1 WHERE type_code = '1'
-- Insert the records for type 2, if the id does not exist for type 1
INSERT INTO #Contacts SELECT id, NULL, phone_num FROM test1 WHERE NOT EXISTS ( SELECT 1 FROM #Contacts WHERE #Contacts.id = test1.id ) AND test1.type_code = '2'
-- if the id has both type 1 and 2, update the phone2 column with the data from type 2
UPDATE #Contacts SET phone2 = test1.phone_num FROM #contacts JOIN test1 ON test1.id = #Contacts.id WHERE type_code = '2' SELECT id, phone1, phone2 FROM #Contacts DROP TABLE #Contacts
How I could accomplish taking several rows for one account and concatenate them into one row, for example I have account_num, invoice_date, transaction_num, msg_counter,Message_2,SQL_LAST_UPDATE the special characters &,",!,$,# are used to determine the Message_2 content for a given account_number that are supposed to be together.
I am needing to put all of that accounts_messages in one row to display on a report, the table I am pulling this data from only has a varchar(40) for the message_2, a proprietary source so can't change that length, "I'VE ASKED THEM TO DO THIS, AND THEY REFUSED". So my only option is to insert this data into my table and create a single Message_2 for that account.
00000000332015-01-16 10:09:43.00000&19 confirmation so 2015-01-19 15:34:59.000 00000000332015-01-16 10:09:43.00000"19ACCT 186743. HE SAID RADIO HAD 2015-01-19 15:34:59.000 00000000332015-01-16 10:09:43.00000!19CALLED Carl ABOUT DEACTIVATION OF RADIO 2015-01-19 15:34:59.000 00000000332015-01-16 10:09:43.00000$19FFERENT ACCT # YEARS AGO, BUT 2015-01-19 15:34:59.000 00000000332015-01-16 10:09:43.00000'19I can cancel the (0.00) billing line on 2015-01-19 15:34:59.000
I want to return all rows in table giftregistryitems with an additional column that holds the sum of column `amount` in table giftregistrypurchases for the respective item in table giftregistryitems.
What I tried, but what returns NULL for purchasedamount:
SELECT (SELECT SUM(amount) from giftregistrypurchases gps where registryid=gi.registryid AND gp.itemid=gps.itemid) as purchasedamount,* FROM giftregistryitems gi LEFT JOIN giftregistrypurchases gp on gp.registryid=gi.id WHERE gi.registryid=2
How can I achieve what I need?
Here's my table definition and data:
/****** Object: Table [dbo].[giftregistryitems] Script Date: 02-05-15 22:37:11 ******/ SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO CREATE TABLE [dbo].[giftregistryitems]( [id] [int] IDENTITY(1,1) NOT NULL,
ColA ColB ----- ----- 21 A 22 A 23 A 24 B 25 B 26 D
What I want is to be able to identify a set sequence (1,2,3) based upon ColB such that I'd get the following result:
ColA ColB ColC ----- ----- ----- 21 A 1 22 A 1 23 A 1 24 B 2 25 B 2 26 D 3
I know that I should be able to get it using ROW_NUMBER() OVER (PARTITION BY ColB ORDER BY ColA), but instead of getting the sequence (1,1,1,2,2,3) I get (1,2,3,1,2,1). Using DENSE_RANK gave me the same results.
I've got a piece of code that returns 53 records when using just the SELECT section.When I change it to INSERT INTO ..... SELECT it only inserts 39 records into the receiving table.There are no keys/contraints/indices or anything else on the receiving table (it's just a dumping ground for some data that will be processed later).
The code for creating the table is here:- USE [CDSExtractInpatients6.2] GO /****** Object: Table [dbo].[CDS_Inpatients_CDS_Feeds_Import] Script Date: 22/05/2015 15:54:15 ******/ SET ANSI_NULLS ON GO
[code]...
I know most of the date fields are being created as varchar on here, but this is something I inherited and the SELECT is outputting the dates as text.Don't know if it makes any difference, but the server is running SQL2008.
I have a query that returns a bunch of rows. I have an sp that takes 3 of the columns in the query as parameters that I want to execute for each row. I can do that easily enough with a cursor, but I thought I'd try to eliminate one more task where I fall back on using cursors.
Is there a efficient way to compare two different columns of 2 different rows in a data set as shown below.
For eg: I would like to DateDiff between Date2 of RowID 1 and Date1 of RowID 2 of IDNo 123. After this comparision , if datediff between two dates are <=14 then i want to update 1 else 0 in IsDateDiffLess14 of RowID1 . In below example its 0 because datediff of two dates >=14. So, want to compare the Date2 and Date1 in this sequence for the same IDNo. For RowID 6 there is only 1 row and no other row to compare, in this case IsDateDiffLess14 should be updated with 0.
I have a query that produces unique rows. However, some of the unique rows have the column called testname that has the same test listed more than once. All I want to do is drop the older testname and keep the testdate column.
Select Distinct TestID, TestDate, TestName
From third.test
I want to keep testdate in the query....guessing I need to put in to a temp table then drop the oldest one somehow by doing a subquery using Select Max....
This is a followup to a previous question to a previous but in reverse of Find rows where value in column not found in another row
Given one table, Table1, with columns Key1 (int), Key2 (int), and Type (varchar)...
I would like to exclude any two rows where Type is equal to 'TypeA' and Key2 is Null that have a corresponding row in the table where Type is equal to 'TypeB' and Key2 is equal to Key1 from another row.
I would like to return all the rows except where Key=1 and Key=3 because those rows together meet the criteria of Type='TypeA'/Key2=NULL and does have a corresponding row with Type='TypeB'/Key1=Key2.
I have a table (represented by #Events) that holds modifications made to another table. I do have some control over the table structure and indexing. I want to pull all of the change records that were made between two dates.
The tricky part is to include the previous version of each record, which will usually be found prior to the start date in question.
The code that I have provided below works. So you can use it to easily see what should be returned. But it's very slow in production.
Any better method to pull this data together?
-- Production version of this table has 4.5 million rows (roughly 1,000 rows per day) -- Primary key is on L4Ident (clustered) -- nonclustered index on ProcessDate, LinkRL4 DROP TABLE dbo.#Events; DROP TABLE dbo.#Results; CREATE TABLE dbo.#Events ( L4Ident int IDENTITY(1,1) NOT NULL,
I have 5 columns in my database. 1 column is coming like a dynamic.
I want to convert records from rows to columns. Currently I have a data like this.
Race AgeRange Amount
W 17-20 500 W 21-30 400 W 31-40 200 A 17-20 100 H 41-50 250 H 51-60 290
So age range is not fixed and it can be any and I have one separate relational table for age range where it's coming from. Now I want to convert it into columns like
I run the script below once a day to keep track of row count over time. I would like to compare the results from today and yesterday to see if anyone deleted more than 20% of data from any given table. How would I do this? I really don't need the data anymore than a day just to compare the results.
Mon - Run script to collect row count Tues - Run script to collect current row into temp table ,compare all row count in both tables ,purge records from Monday and insert current Wed - Run script to collect current row into temp table ,compare all row count in both tables
What I need to be able to find is any records where the Discontinue_Date is greater than the Effective_Date on the next row for a given Customer ID and Part_ID. This is a customer pricing table so the Discontinue_Date of row 53 for example should never be greater than the Effective_Date of row 54130, these are the records I'm looking to find. So I'm looking for a SELECT query that would look for any records where this is true. Obviously the last Discontinue_Date row for a Customer_ID will not have a next row so I wouldn't want to return that.
I am needing to combine the Notes field where Number and date are the same...For example
for Number 0000000003 I need notes to Read ('CHK # 2452 FOR $122.49 REJECTED AS NSF ON 2/25/15') the note counter is different for each row, and is combination of special char, 0-Z and looks like the (!) depicts the start of a new Number.
I currently have data stored in a temporary table and I would like to transpose the data into a better format. I would like for the query to be dynamic since one of the tables currently has over 500 columns.
The attached file provides an example of the table structure along with sample data. Below the first set of data is the desired final format.
