How to achieve this result using a T-SQL select query.
Given this sample table :
create table sample (a int, b int)
insert into sample values (999, 10)
insert into sample values (16, 11)
insert into sample values (10, 12)
insert into sample values (25, 13)
insert into sample values (999, 20)
insert into sample values (14, 12)
insert into sample values (90, 45)
insert into sample values (18, 34)
I'm trying to achieve this output:
a b result
----------- ----------- -----------
999 10 10
16 11 10
10 12 10
25 13 10
999 20 20
14 12 20
90 45 20
18 34 20
The rule is fairly simple: if column 'a' has the special value of 999 the result for that row and following rows (unless the value of 'a' is again 999) will be the value of column 'b'. Assume the first record will have 999 on column 'a'.
Any hint how to implement, if possible, the select query without using a stored procedure or function?
Thank you.
António
You can do what you want if you add a column to specify the ordering:
create table sample (
id int identity(1, 1),
a int,
b int
);
Then you can do what you want by finding the "999" version that is most recent and copying that value. Here is a method using window functions:
select a, b, max(case when a = 999 then b end) over (partition by id_999) as result
from (select s.*,
max(case when a = 999 then id end) over (order by id) as id_999
from sample s
) s;
You need to have an id column
select cn.id, cn.a
, (select top (1) b from sample where sample.id <= cn.id and a = 999 order by id desc)
from sample as cn
order by id
Related
Here are the important columns in my table
ItemId RowID CalculatedNum
1 1 3
1 2 0
1 3 5
1 4 25
1 5 0
1 6 8
1 7 14
1 8 2
.....
The rowID increments to 141 before the ItemID increments to 2. This cycle repeats for about 122 million rows.
I need to SUM the CalculatedNum field in groups of 6. So sum 1-6, then 7-12, etc. I know I end up with an odd number at the end. I can discard the last three rows (numbers 139, 140 and 141). I need it to start the SUM cycle again when I get to the next ItemID.
I know I need to group by the ItemID but I am having trouble trying to figure out how to get SQL to SUM just 6 CalculatedNum's at a time. Everything else I have come across SUMs based on a column where the values are the same.
I did find something on Microsoft's site that used the ROW_NUMBER function but I couldn't quite make sense of it. Please let me know if this question is not clear.
Thank you
You need to group by (RowId - 1) / 6 and ItemId. Like this:
drop table if exists dbo.Items;
create table dbo.Items (
ItemId int
, RowId int
, CalculatedNum int
);
insert into dbo.Items (ItemId, RowId, CalculatedNum)
values (1, 1, 3), (1, 2, 0), (1, 3, 5), (1, 4, 25)
, (1, 5, 0), (1, 6, 8), (1, 7, 14), (1, 8, 2);
select
tt.ItemId
, sum(tt.CalculatedNum) as CalcSum
from (
select
*
, (t.RowId - 1) / 6 as Grp
from dbo.Items t
) tt
group by tt.ItemId, tt.Grp
You could use integer division and group by.
SELECT ItemId, (RowId-1)/6 as Batch, sum(CalculatedNum)
FROM your_table GROUP BY ItemId, Batch
To discard incomplete batches:
SELECT ItemId, (RowId-1)/6 as Batch, sum(CalculatedNum), count(*) as Cnt
FROM your_table GROUP BY ItemId, Batch HAVING Cnt = 6
EDIT: Fix an off by one error.
To ensure you're querying 6 rows at a time you can try to use the modulo function : https://technet.microsoft.com/fr-fr/library/ms173482(v=sql.110).aspx
Hope this can help.
Thanks everyone. This was really helpful.
Here is what we ended up with.
SELECT ItemID, MIN(RowID) AS StartingRow, SUM(CalculatedNum)
FROM dbo.table
GROUP BY ItemID, (RowID - 1) / 6
ORDER BY ItemID, StartingRow
I am not sure why it did not like the integer division in the select statement but I checked the results against a sample of the data and the math is correct.
I have the following table:
No Item Value
----------------------------
1 A 5
2 B 8
3 C 9
If I use Min function on Value field, then I'll get 5.
My question is, how can I put the MIN value into a new column? Like the following result:
No Item Value newCol
----------------------------
1 A 5 5
2 B 8 5
3 C 9 5
Is it possible to do that?
Thank you.
Something like:
select No, Item, Value, (select min(value) from table)
from table
should do it.
I'd prefer to do the subquery in a join, you'll have to name the field. Something like this;
Sample Data
CREATE TABLE #TestData (No int, item nvarchar(1), value int)
INSERT INTO #TestData (No, item, value)
VALUES
(1,'A',5)
,(2,'B',8)
,(3,'C',9)
Query
SELECT
td.No
,td.item
,td.value
,a.Min_Value
FROM #TestData td
CROSS JOIN
(
SELECT
MIN(Value) Min_Value
FROM #TestData
) a
Result
No item value Min_Value
1 A 5 5
2 B 8 5
3 C 9 5
You could do that even simpler by using an appropriate OVER() clause.
SELECT *
, MIN(Value) OVER () AS [newCol]
FROM Table
This would be simpler and less resource consuming than a (SELECT MIN(Value) FROM TABLE) in the top level SELECT.
Sample code:
DECLARE #tbl TABLE (No int, Item char(1), Value int)
INSERT #tbl VALUES (1, 'A', 5), (2, 'B', 8), (3, 'C', 9)
SELECT *
, MIN(Value) OVER () AS [newCol]
FROM #tbl
Using cross join with min value from table :
SELECT * FROM #Tbl1 CROSS JOIN (SELECT MIN(Value1) Value1 FROM #Tbl1) A
Suppose there is a table with 3 rows for employee A and 2 rows for employee B. I need to write a query which will return exactly 3 rows for both the employee (i.e. a dummy row should be present for employee B)
Your question is not completely clear to me (for instance, can "3 rows" be "8 rows" too), but some starting points:
Set up
create table emprec
( emp_id number
, val number
)
insert into emprec values ( 1, 15)
insert into emprec values ( 1, 16)
insert into emprec values ( 1, 17)
insert into emprec values ( 2, 18)
insert into emprec values ( 2, 19)
Query
select driving_x_axis.counter
, driving_y_axis.emp_id
, emprec.val
from ( select level counter
from dual
connect by level <= 3
) driving_x_axis
join ( select distinct emp_id
from emprec
) driving_y_axis
on 1=1 /* Carthesian. */
left
outer
join ( select emp_id
, val
, row_number() over (partition by emp_id order by val) rownumber
from emprec
) emprec
on emprec.rownumber = driving_x_axis.counter
and emprec.emp_id = driving_y_axis.emp_id
Result
1 1 15
2 1 16
3 1 17
1 2 18
2 2 19
3 2 <null>
Simpler alternatives may be possible; please elaborate on your question.
I've got a query I'm working on and I want to increment one of the fields and restart the counter when a key value is different.
I know this code doesn't work. Programmatically this is what I want...
declare #counter int, #id
set #counter = 0
set #id = 0
select distinct
id,
counter = when id = #id
then #counter += 1
else #id = id
#counter = 1
...with the end result looking something like this:
ID Counter
3 1
3 2
3 3
3 4
6 1
6 2
6 3
7 1
And yes, I am stuck with SQL2k. Otherwise that row_number() would work.
Assuming a table:
CREATE TABLE [SomeTable] (
[id] INTEGER,
[order] INTEGER,
PRIMARY KEY ([id], [order])
);
One way to get this in Microsoft SQL Server 2000 is to use a subquery to count the rows with the same id and a lower ordering.
SELECT *, (SELECT COUNT(*) FROM [SomeTable] counter
WHERE t.id = counter.id AND t.order < counter.order) AS row_num
FROM [SomeTable] t
Tip: It's 2010. Soon your SQL Server will be old enough to drive.
If you use SQL Server 2005 or later, you get wonderful new functions like ROW_NUMBER() OVER (PARTITION...).
Yes you want ROW_NUMBER().
I would try:
SELECT id, ROW_NUMBER() OVER (PARTITION BY ID ORDER BY ID) AS Counter
One way to do this is to throw the data into a temp table with an identity column that is used as a row number. Then make the counter column a count of the other rows with the same Id and a lower row number + 1.
CREATE TABLE #MyData(
Id INT
);
INSERT INTO #MyData VALUES(3);
INSERT INTO #MyData VALUES(3);
INSERT INTO #MyData VALUES(3);
INSERT INTO #MyData VALUES(3);
INSERT INTO #MyData VALUES(6);
INSERT INTO #MyData VALUES(6);
INSERT INTO #MyData VALUES(6);
INSERT INTO #MyData VALUES(7);
CREATE TABLE #MyTempTable(
RowNum INT IDENTITY(1,1),
Id INT,
Counter INT
);
INSERT INTO #MyTempTable
SELECT Id, 0
FROM #MyData
ORDER BY Id;
SELECT Id, (SELECT COUNT(*) + 1 FROM #MyTempTable WHERE Id = t1.Id AND RowNum < t1.RowNum) AS 'Counter'
FROM #MyTempTable t1;
You should get the following output based on your example:
Id Counter
3 1
3 2
3 3
3 4
6 1
6 2
6 3
7 1
Having row_number() means you have to deal with far, far fewer correlated subqueries. #Bill Karwin's solution works (+1); here's another version that does the same thing but that might be a bit easier to follow. (I used datetimes to determine ordering.)
-- Test table
CREATE TABLE Test
( Id int not null
,Loaded datetime not null
)
-- Load dummy data with made-up distinct datetimes
INSERT Test values (3, 'Jan 1, 2010')
INSERT Test values (3, 'Jan 2, 2010')
INSERT Test values (3, 'Jan 5, 2010')
INSERT Test values (3, 'Jan 7, 2010')
INSERT Test values (6, 'Feb 1, 2010')
INSERT Test values (6, 'Feb 11, 2010')
INSERT Test values (7, 'Mar 31, 2010')
-- The query
SELECT t1.Id, count(*) Counter
from Test t1
inner join Test t2
on t2.Id = t1.Id
and t2.Loaded <= t1.Loaded
group by t1.Id, t1.Loaded
-- Clean up when done
DROP TABLE Test
It is important to note that, without good indexes (and perhaps even with them), these kinds of queries can perform very poorly, particularly on large tables. Check and optimize carefully!
For MySql, I was able to make it with this query.
SELECT (SELECT COUNT(id) +1 FROM sku s WHERE t.item_id = s.item AND s.id < t.sku_id) AS rowNumber, t.*
FROM
(select item.Name as itemName ,item.id as item_id , sku.Name as skuName ,sku.id as sku_id from item
INNER JOIN sku ON item.id = sku.item
WHERE item.active = 'Y'
) t
1 Roasted Pistachios (Salted, In Shell) 84 1 Pound Bags 84
3 Roasted Pistachios (Salted, In Shell) 84 25 Pound Cases 1174
5 Roasted Pistachios (Salted, In Shell) 84 12 x 2.6 Ounce Bags 5807
2 Roasted Pistachios (Salted, In Shell) 84 5 Pound Bags 814
4 Roasted Pistachios (Salted, In Shell) 84 Samples 4724
6 Roasted Pistachios (Salted, In Shell) 84 12 x 3.2 Ounce Bags 18145
4 Star Fruit 981 5 Pound Bags 17462
1 Star Fruit 981 1 Pound Bags 2125
3 Star Fruit 981 11 Pound Bags 2226
2 Star Fruit 981 44 Pound Cases 2156
I've received data from an external source, which is in a summarised format. I need a way to disaggregate this to fit into a system I am using.
To illustrate, suppose the data I received looks like this:
receivedTable:
Age Gender Count
40 M 3
41 M 2
I want this is a disaggregated format like this:
systemTable:
ID Age Gender
1 40 M
2 40 M
3 40 M
4 41 M
5 41 M
Thanks
Karl
Depending of the range of your count you could use a lookup table that holds exactly x records for each integer x. Like this:
create table counter(num int)
insert into counter select 1
insert into counter select 2
insert into counter select 2
insert into counter select 3
insert into counter select 3
insert into counter select 3
insert into counter select 4
insert into counter select 4
insert into counter select 4
insert into counter select 4
then join with this table:
create table source(age int, gender char(1), num int)
insert into source select 40, 'm', 3
insert into source select 30, 'f', 2
insert into source select 20, 'm', 1
--insert into destination(age, gender)
select age, gender
from source
inner join counter on counter.num = source.num
From the "Works on my machine (TM)" stable a recursive query, with all the usual caveats about maximum recursion depth.
with Expanded(exAge, exGender, exRowIndex) as
(
select
Age as exAge,
Gender as exGender,
1 as exRowIndex
from
tblTest1
union all
select
exAge,
exGender,
exRowIndex+1
from
tblTest1 t1
inner join
Expanded e on (e.exAge = t1.Age and e.exGender = t1.Gender and e.exRowIndex < t1.Count)
)
select
exAge,
exGender,
exRowIndex
from
Expanded
order by
exAge,
exGender,
exRowIndex
option (MAXRECURSION 0) -- BE CAREFUL!!
You don't get the row identifier - but inserting the result of the query into a table with an identity column would deal with that.