Problem With Where Clause When Multiple DESC Indexes
May 12, 2007
I have come across a problem executing a select with a multi-part where clause that only shows up if there are multiple indexes on the table. The situation using a simplified table is shown below
create table tblTest( utcTimestamp datetime NOT NULL, testType int NOT NULL)go
insert into tblTest (utcTimestamp, testType) VALUES('6/1/2003 0:0:0', 100)go
Now, without adding any indexes to the table, I can execute the following select and it works fine, returning the single row in 2003:
select * from tblTest where utcTimestamp < '1/1/2004 0:0:0' and utcTimestamp > '1/1/2003 0:0:0' and testType = 100go
Furthermore, if I introduce a single descending index on just the utcTimestamp:
CREATE INDEX IX_tblTest_Timestamp ON tblTest (utcTimestamp DESC)go
the search still works.
HOWEVER, if I now introduce another index:
CREATE INDEX IX_tblTest_EntryType_Timestamp ON tblTest ( testType, utcTimestamp DESC)go
the search does **not** return the row.
However, if I change the where clause to remove the test of testType:
select * from tblTest where utcTimestamp < '1/1/2004 0:0:0' and utcTimestamp > '1/1/2003 0:0:0'go
it works.
Also, strangely, if I populate the table with a number of records with different dates and execute the following search:
select * from tblTest where utcTimestamp > '1/1/2004 0:0:0' and testType = 100go
I get records from **earlier** than 1/1/2004 (i.e. like the sense of the compare is wrong)
Finally, as I was writing this report, I discovered that all of these problems go away if the DESC is removed from the indexes - so that's my workaround, but it still looks like a bug.
Okay this is a test...actually I am still learning SQL and need some help. Does anyone have any information on being able to move indexes from one database to another. My scenario is I have 3 databases, Development, QA and Production. I want to move/copy indexes I created in Development to the QA database. I have many indexes so I do not want to have to recreated them if I can avoid it. Any suggestions?
where exigo_data_sync.orderdetail.itemcode in (B1001, B1001B, B1007, B1007B, B1008, B1008B, B1000, B1000B, B1006, B1006B, B1009, B1009B)
I keep getting a ADO error stating invalid column names...these are not column names they are the data that i want to use in the where clause. What am I doing wrong?
I work in the healthcare area, and am handling the survey data ETL's. There are around 8 different survey areas and based on information received from them for the visit they reference, I want to pull in more info from our invoicing database. My idea is this:
1.) Pull in the flat file to an ODBC staging table 2.) Cache all invoice records that fall between the MIN(Date of Service) and MAX(Date of Service) from the staging table. 3.) First lookup the information needed on patientID, providerID, date of service, and billing location. 4.) For the surveys that didn't match on those 4 columns, try looking up based on patientID, date of service, and billing location (since I could be 99% sure this would still return the record I need). 5.) For the remaining surveys, lookup based just on patientID and date of service. These records will be flagged for manual review because clearly, if a patient has multiple appointments in the same day, this will be prone to error.
However, in trying to use only 3 of the columns in the lookup, I get the error saying basically that I need to utilize all 4. Is there a way around this, or is there an entirely different way I should be approaching this? The reason I thought cache transform was the answer is because I will need to run a different package for each lookup, as the data and logic between each survey will vary, but the invoice data "pool" will stay the same regardless.
hello. I have a database that a client developed that I need to pull data from. it consists of articles that fall into a range of 3 main categories. each article will have up to 7 different subcategories they fall into. I need to be able to sort by main category as well as by subcategory. But when I create the SQL query it gets really messy. I tried using WHERE Cat1= comm OR leg OR and so on, but there are seven categories so this gets very cumbersome and doesn't quite work. Is there a way to create an array or a subquery for this? I am a total newbie, so any help is much appreciated!
How can you handle multiple criteria query in T-SQL ? i wrote selection query and in my where clause i have about 7 different criteria and for some reason when i run the query i do not get any error but i do not get any data return.
So is there any other way to handle multiple criteria in T-SQL ?
I have a table with a field that contains an integer which represents the state of a record. This field "intType" may contain values 0-4.
A parameter in my stored procedure "@intUserType" may contain values 0-3
If @intUserType = 0, I need to select the records where intType = 0 or 3 but if @intUserType = 3, I need to return all records where intType > 1, all other values of @intUserType should return no records
The query I am working with seems a bit forced and I feel like it could be simplified, but I can't seem to wrap my head around it.
This is what I am working with:
Code: SELECT * FROM tblEmployees WHERE (intType = (CASE WHEN @intUserType = 0 THEN 0 ELSE NULL END) OR intType = (CASE WHEN @intUserType = 0 THEN 3 ELSE NULL END) OR intType > (CASE WHEN @intUserType = 3 THEN 1 ELSE NULL END))
Maybe it is as good as it needs to be ... I don't know .. I've only been using SQL regulary for a couple of months and I have not had the time to really study it in depth.
I need to return all the distinct IDs where the combination of AttrID and AttrVal matches multiple criteria for that specific ID.
I have the following:
SELECT DISTINCT(ID) FROM ((SELECT a.ID FROM tblAttributes AS a WHERE a.AttrID = 90 AND a.AttrVal = 7) AS x INNER JOIN (SELECT a.ID FROM tblAttributes AS a WHERE a.AttrID = 91 AND a.AttrVal = 8) AS y ON x.ID = y.ID)
Hello People, Please help. I have a basic report with a parameter in the 'Where" clause called (@Stat) from the statement below: " WHERE contractinfo.termdate >= GETDATE() AND provider.status= 'Active' AND provider.credentialstatus = (@Stat)"
This variable has one of two values: 'A' or 'B' that the user selects, how do I set it up so that if user selects say 'A' then the Where clause would go to one set of constraints ie "WHERE contract.description NOT LIKE 'NON%' "
But if the user selects 'B' then the Where clause would go to a different set of constraint ie "WHERE contract.description LIKE 'NON%' " Thanks
Hi AllI am having a problem with an ORDER BY clause when selecting information from multiple tables. EgSELECT i.InvoiceId, pd.PayDescription, u.UserNameFROM Invoice i LEFT OUTER JOIN tblPay ON i.PayId = pd.PayId LEFT OUTER JOIN tblUsers ON i.UserId = u.UserIdORDER BY pd.PayDescriptionthis is just an example my query is a lot more complex. Is there any simply way you can do an order by in this way?I am writing this for MSSQL Server 2000ThanksBraiden
HiI'm a bit stuck with a SELECT query. This is a simplified version ofwhat I need. I've had a look in a few books and online but I'mdefinitely missing something. I'm trying to avoid looping and cursors.I'll be running this in a stored procedure on SQL 7.I have a separate query which returns a series of numbers, A, say 101103 107 109 113.I have a table (tableB) with a field myFieldB where I have anotherseries of numbers, B. I want return each row in tableB wherei - ALL values in A existii- ANY values in A existFor ii, I can use WHERE myFieldB IN AHow about for i?Is there a good guide on the web or a book on WHERE clauses and/ormore complex SQL?Thanks in advance!Sam
I have a parent and child package. i pass a parent package variable called @abc with a value of (1,2,3,4,5,6) to the child package. here in the oledb source i have a select statement like, select * from A where id in (@abc)
Right now I have to do something like this and it is time consuming every time I have to query a specific table...
SELECT lots_of_columns FROM table WHERE (column5 = '1' OR column6 = '1' OR column7 = '1' OR column8 = '1' OR column9 = '1' OR column10 = '1' OR column11 = '1' OR column12 = '1') AND other_query_critiera_here
Typing out the OR statement gets long, time consuming and prone to errors because that first where line with all the ORs can sometimes have 20+ ORs in it. As some insight, the columns are text columns, sometimes they have data, sometimes they are NULL. Sometimes they have the same data (i.e., column5 and column6 and column12 could both have '1' as values).
I created a stored procedure like the following in the hope that I can pass mulitple company_id to the select statement:
CREATE PROC sp_test @in_company_code nvarchar(1024) AS
select company_code, name, description from member_company where company_code in (@in_company_code)
However, I tried the following :
exec sp_test 'abc', 'rrd', 'bbc'
Procedure or function sp_test has too many arguments specified.
and SQLServer doesn't like it.
Did I specify this stored procedure correct? If so, how can I can pass multiple values to the stored procedure then to the sql statement? If not, is it possible to specify a stored procedure like this?
Hello,I'm using a shape query, but instead of using a simple clause "RELATEfield1 to field2" (relates the parent to the child), i wan't to use 2relates. somthing like "RELATE field1 to field2 AND field3 to field4".I want to receive in the children RS only the records who apply bothconditions.How do i do that ?Thanks !
I would like to know how i can handle multiple columns returned by a subquery via IN clause in case of sql server 2005. I can do that in oracle by using the following statement:
DELETE FROM TEST1 WHERE (ID, ID1) NOT IN (SELECT ID,ID1 FROM TEST2);
Hey guys, I have a view with dates (TheDate) meant to be arranged in descending order. When I 'Execute SQL' while in the view, the DESC order works just fine and shows up with the latest date first going down. However, once I 'OPEN VIEW' the order keeps defaulting to ASCending order.
How do I keep it in DESC order for viewing? Here's the statement:
SELECT TOP (100) PERCENT TheDate FROM dbo.MyDates ORDER BY TheDate DESC
In the followinf portion of my SQL Statement, I cannot figure out how to specify ASC or DESC for the OrderBy... If I Put OrderDate ASC after the THEN, I get an error and if I put @SortDir after the END I get an error... Can Anyone help with this? WHERE OrderID IN(SELECT OrderID FROM #rsltTable WHERE ID >= @l_FirstRecord AND ID <= @l_LastRecord)ORDER BYCASE @OrderByWHEN 'OrderDate ASC' THEN OrderDateWHEN 'OrderDate DESC' THEN OrderDateEND Thank You,Jason
hi, i have headers at the top of my list and would like for people to click the pubs link and it orders the pubs alphabetically descending and when they click again to ascend also to do this with towns and addresses and postcodes?!
Does anyone have a general rule or guide on when to use this SQL 2000 option when creating indexes? I was thinking generally on nonclustered indexes where the column would be unique and incremental and usually filtered on by range and often used in the order by clause. Such as columns of datetime or integers datatypes. Thanks.