I'm trying to write a function that adds values that are selected from two different tables. I need the SUM of one column for a specific id in table 1, added to the SUM of a column for a specific id in table 2.
CREATE FUNCTION dbo.getSum(#id varchar(9)) --Our IDs are strings of length 9
RETURNS integer --I've also tried decimal(x,x) and int
AS
BEGIN
DECLARE #total as integer; -- same here with decimal and int
SELECT #total =
(
(select SUM(Amount)
from table1
where id = #id)
+
(select SUM(amount)
from table2
where id = #id)
);
RETURN #total;
END;
I get several types of errors when I try to create this function, like incorrect syntax near 'integer', Must declare the scalar variable "#total"., and Incorrect syntax near 'END'.
I've tried several variations and looked at several SO questions, but haven't found one that's fixed this issue for me yet. This is on SQL Server.
As previously mentioned the cause of your error is you did not define a datatype for your inbound parameter.
I would suggest a bit of a departure from the current structure. Instead of a scalar function I would use an inline table valued function. Something along these lines.
CREATE FUNCTION dbo.getSum(#id int)
RETURNS table
AS RETURN
SELECT SUM(MyAmount)
from
(
select SUM(Amount) as MyAmount
from table1
where id = #id
UNION ALL
select SUM(amount)
from table2
where id = #id
) x;
Looks like you are missing the parameter type in the function definiton
Try with the following
CREATE FUNCTION dbo.getSum(#id int)
RETURNS integer --I've also tried decimal(x,x)
AS
BEGIN
DECLARE #total as integer; -- same here with decimal
SELECT #total =
(
(select SUM(Amount)
from table1
where id = #id)
+
(select SUM(amount)
from table2
where id = #id)
);
RETURN #total;
END;
I think you need to declare the type of the parameter
CREATE FUNCTION dbo.getSum(#id int)
Your data types should be INT and there are general syntax errors... see below.
CREATE FUNCTION dbo.getSum(#id AS INT) -- Declare your paramater AS (datatype)
RETURNS INT -- No need for AS here, just start your block
BEGIN
DECLARE #total AS INT;
SELECT #total =
(
(select SUM(Amount)
from TableOne
where id = #id)
+
(select SUM(amount)
from TableTwo
where id = #id)
);
RETURN #total;
END;
Related
I get batches of inventory items to update and I would like to eliminate calling the stored procedure multiple times and instead call it once with multiple values. I have done similar in oracle with the parameters as an array trick. I would like to do something similar for SQL Server.
I have a comma separated list of Sku
I have a comma separated list of Quantity.
I have a comma separated list of StoreIds.
The standard update is
Update Inventory
set quantity = #Quantity
where sku = #Sku and StoreId = #StoreId;
Table definition
CREATE TABLE Inventory
(
[Sku] NVARCHAR(50) NOT NULL,
[Quantity] DECIMAL NULL DEFAULT 0.0,
[StoreId] INT NOT NULL
}
My bad attempt at doing this
ALTER PROCEDURE UpdateList
(#Sku varchar(max),
#Quantity varchar(max),
#StoreId varchar(max))
AS
BEGIN
DECLARE #n int = 0;
DECLARE #skuTable TABLE = SELECT CONVERT(value) FROM STRING_SPLIT(#Sku, ',');
DECLARE #quantityTable = SELECT CONVERT(value) FROM STRING_SPLIT(#Quantity, ',');
DECLARE #StoreIdTable = SELECT CONVERT(value) FROM STRING_SPLIT(#StoreId , ',');
WHILE #n < #skuTable.Count
BEGIN
UPDATE inventoryItem
SET Quantity = #quantityTable
WHERE Sku = #skuTable AND StoreId = #StoreIdTable;
SELECT #n = #n + 1;
END
END
I am open to using temp tables as parameters instead of comma separated. This is being called from an Entity Framework 6 context object from the front end system.
It's a bad practice to pass tabular values in this way.
Best solution is to pass it as a "user defined table type", if possible,
otherwise, it's better to get JSON/XML parameter
and then you can update your table like this:
--[ Parameters ]--
DECLARE #json AS NVARCHAR(MAX) = '[{"Sku":"A","Quantity":1.4,"StoreId":1},{"Sku":"B","Quantity":2.5,"StoreId":2},{"Sku":"C","Quantity":3.6,"StoreId":3}]';
--[ Bulk Update ]--
UPDATE inventoryItem
SET Quantity = I.Quantity
FROM inventoryItem AS T
JOIN OPENJSON(#json) WITH (Sku NVARCHAR(50), Quantity DECIMAL(5,1),StoreId INT) AS I
ON I.Sku = T.Sku
AND I.StoreId = T.StoreId
It's a bad practice to pass tabular values as varchar columns parameters,
but if you still want to go this way, here is a working code:
--[ Parameters ]--
DECLARE #Sku VARCHAR(max) = 'A,B,C',
#Quantity VARCHAR(max) = '1.4,2.5,3.6',
#StoreId VARCHAR(max) = '1,2,3'
--[ Converting VARCHAR Parameters to Table #Inventory ]--
DROP TABLE IF EXISTS #Sku
SELECT IDENTITY(int, 1,1) AS RowNum,
T.value
INTO #Sku
FROM STRING_SPLIT(#Sku, ',') AS T
DROP TABLE IF EXISTS #Quantity
SELECT IDENTITY(int, 1,1) AS RowNum,
T.value
INTO #Quantity
FROM STRING_SPLIT(#Quantity, ',') AS T
DROP TABLE IF EXISTS #StoreId
SELECT IDENTITY(int, 1,1) AS RowNum,
T.value
INTO #StoreId
FROM STRING_SPLIT(#StoreId, ',') AS T
DROP TABLE IF EXISTS #Inventory
SELECT Sku.value AS Sku,
Quantity.value AS Quantity,
StoreId.value AS StoreId
INTO #Inventory
FROM #Sku AS Sku
JOIN #Quantity AS Quantity ON Quantity.RowNum = Sku.RowNum
JOIN #StoreId AS StoreId ON StoreId.RowNum = Sku.RowNum
--[ Bulk Update ]--
UPDATE inventoryItem
SET Quantity = I.Quantity
FROM inventoryItem AS T
JOIN #Inventory AS I
ON I.Sku = T.Sku
AND I.StoreId = T.StoreId
The above answers are correct for updates and answered my question. But I wanted to add the insert here as I am sure someone will be looking for both. Maybe I will come back an make a new question and answer it myself.
I think the JSON version is best for my issue because I am doing entity framework and serializing an object to JSON is a trivial task. The basic process is to create an inline temp table from the json string. Calling out the objects via a simple dot notation string. I would suggest making the object passed in as simple as possible and preferably one level of properties.
create or alter Procedure bulkInventoryInsert( #json AS NVARCHAR(MAX))
AS
BEGIN
INSERT into inventory
SELECT Sku, Quantity, StoreId FROM
OPENJSON(#json)
WITH(Sku varchar(200) '$.Sku',
Quantity decimal(5,1) '$.Quantity',
StoreId INT '$.StoreId');
END
DECLARE #json AS NVARCHAR(MAX) = '[{"Sku":"A","Quantity":1.4,"StoreId":2},{"Sku":"B","Quantity":2.5,"StoreId":3},{"Sku":"C","Quantity":3.6,"StoreId":2}]';
EXECUTE bulkInventoryInsert #json;
The key part to recognize is this section here:
SELECT Sku, Quantity, StoreId FROM
OPENJSON(#json)
WITH(Sku varchar(200) '$.Sku',
Quantity decimal(5,1) '$.Quantity',
StoreId INT '$.StoreId');
This is creating a temp table with columns that match the table that it will be inserted into. The "WITH" portion specifies the column name, type, and where in the Json string to get the value.
I hope this will help. Maybe when I get time I will do a question and answer for this.
I have a table [Order] that has records with sequential ID (in odd number only, i.e. 1,3,5,7...989, 991, 993, 995, 997, 999), it is seen that a few records were accidentally deleted and should be inserted back, first thing is to find out what records are missing in the current table, there are hundreds of records in this table
Don't know how to write the query, can anyone kindly help, please?
I am thinking if I have to write a stored procedure or function but would be better if I can avoid them for environment reasons.
Below peuso code is what I am thinking:
set #MaxValue = Max(numberfield)
set #TestValue = 1
open cursor on recordset ordered by numberfield
foreach numberfield
while (numberfield != #testvalue) and (#testvalue < #MaxValue) then
Insert #testvalue into #temp table
set #testvalue = #textvalue + 2
Next
Next
UPDATE:
Expected result:
Order ID = 7 should be picked up as the only missing record.
Update 2:
If I use
WHERE
o.id IS NULL;
It returns nothing:
Since I didn't get a response from you, in the comments, I've altered the script for you to fill in accordingly:
declare #id int
declare #maxid int
set #id = 1
select #maxid = max([Your ID Column Name]) from [Your Table Name]
declare #IDseq table (id int)
while #id < #maxid --whatever you max is
begin
insert into #IDseq values(#id)
set #id = #id + 1
end
select
s.id
from #IDseq s
left join [Your Table Name] t on s.id = t.[Your ID Column Name]
where t.[Your ID Column Name] is null
Where you see [Your ID Column Name], replace everything with your column name and the same goes for [Your Table Name].
I'm sure this will give you the results you seek.
We can try joining to a number table, which contains all the odd numbers which you might expect to appear in your own table.
DECLARE #start int = 1
DECLARE #end int = 1000
WITH cte AS (
SELECT #start num
UNION ALL
SELECT num + 2 FROM cte WHERE num < #end
)
SELECT num
FROM cte t
LEFT JOIN [Order] o
ON t.num = o.numberfield
WHERE
o.numberfield IS NULL;
I'm developing this stored procedure on SQL Server 2012.
The stored procedure will update Quantity rows in EXTERNAL_CODES table for each row in #newBatches parameter. It's like a loop, I will need to create a new row in BATCHES table for each row in #newBatches parameter.
And then, I have to update Quantity rows in EXTERNAL_CODES table with each batchId created.
CREATE PROCEDURE [dbo].[CreateBatchAndKeepExternalCodes]
#newBatches as dbo.CreateBatchList READONLY,
#productId int
AS
set nocount on;
declare #lowestCodeLevel tinyint;
-- ======== VALIDATION ==========
if ((select count(name) from #newBatches) = 0)
return -112;
-- ====== CODE ========
-- Get lowest aggregation level.
set #lowestCodeLevel =
(select min(c.application_code)
from CHINA_CODES_HEADER c, PRODUCTS p
where p.Id = #productId and c.DRUG_TEN_SEATS = p.PRODUCT_CODE);
begin transaction;
insert into BATCHES (PRODUCT_ID, NAME, CREATED)
select #productId, Name, CAST(SYSDATETIMEOFFSET() as nvarchar(50))
from #newBatches;
update top(t.Quantity) EXTERNAL_CODES
set BATCH_ID = (select ID from BATCHES where NAME = t.Name)
, USED = 1
from (select Name, Quantity from #newBatches) t
where PRODUCT_ID = #productId and CODE_LEVEL = #lowestCodeLevel;
commit transaction;
RETURN 0
I get an error on this update:
update top(t.Quantity) EXTERNAL_CODES
set BATCH_ID = (select ID from BATCHES where NAME = t.Name)
, USED = 1
from (select Name, Quantity from #newBatches) t
where PRODUCT_ID = #productId and CODE_LEVEL = #lowestCodeLevel;
The error is here: update top(t.Quantity). It can't find t.Quantity.
dbo.CreateBatchList is:
CREATE TYPE [dbo].[CreateBatchList] AS TABLE
(
Name nVARCHAR(20),
Quantity int
)
My problem is that I can't set to update Quantity rows. Any idea?
The error (or warning) message is:
SQL71005: The reference to the column t.Quantity could not be resolved.
Maybe I could use MERGE.
Your update statement is quite confusing. If for instance #newBatches table has multiple lines, then you are saying, pick all the Quantity from #newBatches in Top?
Anyway, I think the solution is to use a loop to use each line from #newBatches to update. I have modified your code in order to test it on my side, and have replaced all the tables with Table Variables. You may find it helpful.
But still without any Order By clause and without knowing actual business logic, I can't say this solution is correct.
DECLARE #productID int;
DECLARE #lowestCodeLevel int;
DECLARE #EXTERNAL_CODES table(BATCH_ID varchar(100), USED bit, PRODUCT_ID int, CODE_LEVEL int);
DECLARE #BATCHES table(ID int, NAME varchar(100));
DECLARE #newBatches table(Name nVARCHAR(20), Quantity int);
-- we don't know at this point whether #newBatches has some column
-- through which we can uniquely identify a row
-- that is why we are creating this new table in which we have Row_ID column
-- through which we can extract each line
DECLARE #newBatchesWithRowID table(Row_ID int not null identity, Name nVarchar(20), Quantity int);
INSERT INTO #newBatchesWithRowID(Name, Quantity)
SELECT Name, Quantity
FROM #newBatches;
DECLARE #prvRow_ID int;
-- loop to iterate in #newBatchesWithRowID table
WHILE(1 = 1)
Begin
DECLARE #row_ID int = NULL;
DECLARE #Name varchar(100);
DECLARE #Quantity int;
SELECT TOP 1 #row_ID = Row_ID
, #Quantity = Quantity
, #Name = Name
FROM #newBatchesWithRowID
WHERE Row_ID > #prvRow_ID OR #prvRow_ID IS NULL
ORDER BY Row_ID;
If #row_ID IS NULL Break;
SET #prvRow_ID = #row_ID;
update top(#Quantity) #EXTERNAL_CODES
set BATCH_ID = (select ID from #BATCHES where NAME = #Name)
, USED = 1
where PRODUCT_ID = #productId and CODE_LEVEL = #lowestCodeLevel;
END
How do I store the select column in a variable?
I tried this, but it throws me an error "Incorrect syntax":
declare #EmpId int
SELECT dbo.Employee.Id as #EmpId FROM dbo.Employee
select #EmpID = ID from dbo.Employee
Or
set #EmpID =(select id from dbo.Employee)
Note that the select query might return more than one value or rows. so you can write a select query that must return one row.
If you would like to add more columns to one variable(MS SQL), there is an option to use table defined variable
DECLARE #sampleTable TABLE(column1 type1)
INSERT INTO #sampleTable
SELECT columnsNumberEqualInsampleTable FROM .. WHERE ..
As table type variable do not exist in Oracle and others, you would have to define it:
DECLARE TYPE type_name IS TABLE OF (column_type | variable%TYPE | table.column%TYPE [NOT NULL] INDEX BY BINARY INTEGER;
-- Then to declare a TABLE variable of this type: variable_name type_name;
-- Assigning values to a TABLE variable: variable_name(n).field_name := 'some text';
-- Where 'n' is the index value
Assuming such a query would return a single row, you could use either
select #EmpId = Id from dbo.Employee
Or
set #EmpId = (select Id from dbo.Employee)
This is how to assign a value to a variable:
SELECT #EmpID = Id
FROM dbo.Employee
However, the above query is returning more than one value. You'll need to add a WHERE clause in order to return a single Id value.
Using SQL Server 2008 Reporting services:
I'm trying to write a report that displays some correlated data so I thought to use a #table variable like so
DECLARE #Results TABLE (Number int
,Name nvarchar(250)
,Total1 money
,Total2 money
)
insert into #Results(Number, Name, Total1)
select number, name, sum(total)
from table1
group by number, name
update #Results
set total2 = total
from
(select number, sum(total) from table2) s
where s.number = number
select from #results
However, Report Builder keeps asking to enter a value for the variable #Results. It this at all possible?
EDIT: As suggested by KM I've used a stored procedure to solve my immediate problem, but the original question still stands: can I use #table variables in Report Builder?
No.
ReportBuilder will
2nd guess you
treats #Results as a parameter
Put all of that in a stored procedure and have report builder call that procedure. If you have many rows to process you might be better off (performance wise) with a #temp table where you create a clustered primary key on Number (or would it be Number+Name, not sure of your example code).
EDIT
you could try to do everything in one SELECT and send that to report builder, this should be the fastest (no temp tables):
select
dt.number, dt.name, dt.total1, s.total2
from (select
number, name, sum(total) AS total1
from table1
group by number, name
) dt
LEFT OUTER JOIN (select
number, sum(total) AS total2
from table2
GROUP BY number --<<OP code didn't have this, but is it needed??
) s ON dt.number=s.number
I've seen this problem as well. It seems SQLRS is a bit case-sensitive. If you ensure that your table variable is declared and referenced everywhere with the same letter case, you will clear up the prompt for parameter.
You can use Table Variables in SSRS dataset query like in my code where I am adding needed "empty" records for keep group footer in fixed postion (sample use pubs database):
DECLARE #NumberOfLines INT
DECLARE #RowsToProcess INT
DECLARE #CurrentRow INT
DECLARE #CurRow INT
DECLARE #cntMax INT
DECLARE #NumberOfRecords INT
DECLARE #SelectedType char(12)
DECLARE #varTable TABLE (# int, type char(12), ord int)
DECLARE #table1 TABLE (type char(12), title varchar(80), ord int )
DECLARE #table2 TABLE (type char(12), title varchar(80), ord int )
INSERT INTO #varTable
SELECT count(type) as '#', type, count(type) FROM titles GROUP BY type ORDER BY type
SELECT #cntMax = max(#) from #varTable
INSERT into #table1 (type, title, ord) SELECT type, N'', 1 FROM titles
INSERT into #table2 (type, title, ord) SELECT type, title, 1 FROM titles
SET #CurrentRow = 0
SET #SelectedType = N''
SET #NumberOfLines = #RowsPerPage
SELECT #RowsToProcess = COUNT(*) from #varTable
WHILE #CurrentRow < #RowsToProcess
BEGIN
SET #CurrentRow = #CurrentRow + 1
SELECT TOP 1 #NumberOfRecords = ord, #SelectedType = type
FROM #varTable WHERE type > #SelectedType
SET #CurRow = 0
WHILE #CurRow < (#NumberOfLines - #NumberOfRecords % #NumberOfLines) % #NumberOfLines
BEGIN
SET #CurRow = #CurRow + 1
INSERT into #table2 (type, title, ord)
SELECT type, '' , 2
FROM #varTable WHERE type = #SelectedType
END
END
SELECT type, title FROM #table2 ORDER BY type ASC, ord ASC, title ASC
Why can't you just UNION the two resultsets?
How about using a table valued function rather than a stored proc?
It's possible, only declare your table with '##'. Example:
DECLARE ##results TABLE (Number int
,Name nvarchar(250)
,Total1 money
,Total2 money
)
insert into ##results (Number, Name, Total1)
select number, name, sum(total)
from table1
group by number, name
update ##results
set total2 = total
from
(select number, sum(total) from table2) s
where s.number = number
select * from ##results