I have a table like this in SQL Server 2012:
DECLARE #test TABLE (dateField datetime,
col1 varchar(10),
col2 varchar(10)
)
insert into #test
select dateadd(day,-1,getdate()), 'Test A', 'Other A'
union
select dateadd(day,-1,getdate()), 'Test E', 'Other E'
union
select getdate(), 'Test B', 'Other B'
union
select dateadd(day,1,getdate()), 'Test C', 'Other C'
union
select dateadd(day,1,getdate()), 'Test A', 'Other C'
union
select dateadd(day,1,getdate()), 'Test D', 'Other B'
I want to retrieve the data in a specific order.
Order should be:
dateField ASC
Col1 ends with A
Col2 ends with B
If datefield is greater than today's date, then it should switch to col1 check.
If there's no col1 that ends with A, then it should switch to col2 check
If there's no col2 that ends with B, no rows should be returned
Any help would be appreciated. I tried doing case statements in SQL. It didn't work as I expected.
You seem a bit confused about if it is filtered or sorted or a combination of that. This does both.
select
*
from (
SELECT
*
,substring(col1,len(col1),1) [LastCharCol1]
,substring(col2,len(col2),2) [LastCharCol2]
FROM #test
) t
where 1=1
AND (
[LastCharCol1]='A'
OR [LastCharCol2]='B'
)
order by CASE WHEN dateField>CONVERT(DATE,GETDATE()) THEN 1 ELSE 0 END DESC,CASE WHEN [LastCharCol1]='A' THEN 1 ELSE 0 END DESC,CASE WHEN [LastCharCol2]='B' THEN 1 ELSE 0 END DESC
If I understood right, you want to filter records:
select *
from #test
where dateField > getdate() and (col1 like '%A' or col2 like '%B')
Related
I would like to change the column values based on below condition. Appreciate your expert advice on this. I need to prefix the respective column value 'F~' if the column is found in col_lst. Can I achieve this through sql query?
Current Output:
row_id|col_lst|col1|col2|col3|col4
1|col1,col2|1|2|3|4
2|col3,col4|A|B|C|D
3|col2|X|Y|Z|P
Expected Output:
row_id|col_lst|col1|col2|col3|col4
1|col1,col2|F~1|F~2|3|4
2|col3,col4|A|B|F~C|F~D
3|col2|X|F~Y|Z|P
Regards
Mithil
You can use the INSTR function on each column name to determine if it is in the col_lst column.
with test_data(row_id,col_lst,col1,col2,col3,col4) as
( select 1, 'col1,col2', '1', '2', '3', '4' from dual union all
select 2, 'col3,col4', 'A', 'B', 'C', 'D' from dual union all
select 3, 'col2', 'X', 'Y', 'Z', 'P' from dual
)
select row_id
, col_lst
, case when instr(col_lst,'col1')>0 then 'F~' || nvl(col1,' ') else col1 end col1
, case when instr(col_lst,'col2')>0 then 'F~' || nvl(col2,' ') else col2 end col2
, case when instr(col_lst,'col3')>0 then 'F~' || nvl(col3,' ') else col3 end col3
, case when instr(col_lst,'col4')>0 then 'F~' || nvl(col4,' ') else col4 end col4
from test_data;
However this is a case where you should just tell the other team NO. There is nothing added that they cannot determine from the col_lst themselves, and it serves no purpose except output formatting - which the presentation layer should be doing.
I have a table with name "PrintWord" and column name as col_letter and data in it is as follows:
"col_letter"
S
A
C
H
I
N
I would like to print the o/p from this table as:
SACHIN
Thanks!
DECLARE #t table
(
Name varchar(10)
)
INSERT INTO #t
SELECT 's' UNION ALL
SELECT 'a' UNION ALL
SELECT 'c' UNION ALL
SELECT 'h' UNION ALL
SELECT 'i' UNION ALL
SELECT 'n'
SELECT DISTINCT
stuff(
(
SELECT ' '+ [Name] FROM #t FOR XML PATH('')
),1,1,'')
FROM (SELECT DISTINCT Name FROM #t ) t
There is a hard-coded version :
SELECT col_letter
FROM PrintWord
ORDER BY
CASE col_letter
WHEN 'S' THEN 1
WHEN 'A' THEN 2
WHEN 'C' THEN 3
WHEN 'H' THEN 4
WHEN 'I' THEN 5
WHEN 'N' THEN 6
END
FOR XML PATH('')
You need an ORDER BY clause to guarantee the order of the letters.
I gained some help from this question, but still need some further assistance.
I need to be able to generate the next available 2-digit alphanumeric code. I cannot change the table definition, before you ask. I am working in T-SQL.
So, for example, let's say I have the sequence
00, 01, 02,..., 09, 0A, 0B, 0C,..., 0Y, 0Z, 10, 11,...1Y, 1Z, 20, 21,..., 9Y, 9Z, I would like for the next id to be A0,
then A1, A2, ..., A9, AA, AB, AC, ..., AZ, I would like for the next id to be B0, then B1, etc.
So, in short, I would like to go from 00 all the way to ZZ and each time look for the MAX in that field and assign a new code 1 greater than the max. I would understand that A > 9, and the first column greater than the second, so A0 > 99 and AA > A9.
I wish I could just assign a numeric id to all of this, but the table definition is more critical at this point and so I'm not allowed to change it, so I am trying to maximize the available ids I'll have in such a limited space.
Thank you for your help.
Have a look at this. This is a really nasty problem for ID's. You've effectively limited yourself a low number of permutations of the key with 2 characters. Also you have a problem that you'll need to deal with if ZZ is used and this algorithm runs again. I have expanded these into as logical steps as possible for demonstration, but feel free to condense as needed.
DECLARE #ExistingTable TABLE (ID CHAR(2))
INSERT INTO #ExistingTable (ID) VALUES ('5A'),('5B')
DECLARE #NewID CHAR(2)
;WITH
Ranks AS (
SELECT '0' AS [Character] UNION SELECT '1' AS [Character] UNION SELECT '2' UNION SELECT '3' UNION SELECT '4' UNION SELECT '5' UNION SELECT '6' UNION
SELECT '7' UNION SELECT '8' UNION SELECT '9' UNION SELECT 'A' UNION SELECT 'B'UNION
SELECT 'C' UNION SELECT 'D' UNION SELECT 'E' UNION SELECT 'F' UNION SELECT 'G' UNION SELECT 'H' UNION
SELECT 'I' UNION SELECT 'J' UNION SELECT 'K' UNION SELECT 'L' UNION SELECT 'M' UNION SELECT 'N' UNION
SELECT 'O' UNION SELECT 'P' UNION SELECT 'Q' UNION SELECT 'R' UNION SELECT 'S' UNION SELECT 'T' UNION
SELECT 'U' UNION SELECT 'V' UNION SELECT 'W' UNION SELECT 'X' UNION SELECT 'Y' UNION SELECT 'Z'
), Permutations AS (
SELECT SecondChar.[Character] + FirstChar.[Character] AS PermuteID
FROM Ranks AS FirstChar
CROSS JOIN Ranks AS SecondChar
), PermutationsKeyed AS (
SELECT ROW_NUMBER() OVER (ORDER BY PermuteID ASC) AS PrimaryKeyHolder,
PermuteID
FROM Permutations
), MaxPK AS (
SELECT MAX(Perm.PrimaryKeyHolder) + 1 AS MaxPK
FROM #ExistingTable AS E
INNER JOIN PermutationsKeyed AS Perm ON (E.ID = Perm.PermuteID)
)
SELECT #NewID = Perm.PermuteID
FROM PermutationsKeyed AS Perm
INNER JOIN MaxPK AS M ON (Perm.PrimaryKeyHolder = M.MaxPK)
SELECT #NewID
I'm not sure how you wanted to go about returning the next value but I think this a simple and efficient ways to get all your values. Let me know if you need anything else.
DECLARE #values TABLE (val CHAR(1));
DECLARE #int INT = 48,
#letters INT = 65;
IF OBJECT_ID('dbo.tbl_keys') IS NOT NULL
DROP TABLE dbo.tbl_keys;
--This will hold the values so you can always reference them
CREATE TABLE dbo.tbl_Keys
(
--Primary key will create a clustered index on rank_id by default
rank_id INT PRIMARY KEY,
ID_Code CHAR(2)
);
--Another index on ID_Code
CREATE NONCLUSTERED INDEX idx_ID_Code ON tbl_keys(ID_Code);
--This is how I get all your individual values
WHILE (SELECT COUNT(*) FROM #values) < 36
BEGIN
IF(#int <= 57)
INSERT INTO #values VALUES(CHAR(#int));
INSERT INTO #values
VALUES (CHAR(#letters))
SET #int = #int + 1;
SET #letters = #letters + 1;
END
--Insert all possible combinations and rank them
INSERT INTO tbl_Keys
--ASCII is your best friend. It returns the ASCII code(numeric value) for characters
SELECT ROW_NUMBER() OVER (ORDER BY ASCII(A.val),ASCII(B.val)) AS rank_id,
A.val + B.val ID
FROM #values A
CROSS JOIN #values B;
I provide two different ways of getting the next ID_code(Read comments):
--Here's some dummy data
WITH CTE_DummyTable
AS
(
SELECT '00' ID_Code
UNION ALL
SELECT '01'
UNION ALL
SELECT '02'
)
----Here's how to get the next value with the assumption there are no gaps in your data
--SELECT MIN(ID_Code) next_id_code
--FROM tbl_Keys
--WHERE ID_code > (SELECT MAX(id_code) FROM CTE_DummyTable)
--This one doesn't assume the gaps and returns the lowest available ID_code
SELECT MIN(ID_Code) next_id_code
FROM tbl_Keys
WHERE ID_code NOT IN (SELECT DISTINCT id_code FROM CTE_DummyTable)
Note: If you were ever to want to convert your alphanumeric values really easily for whatever reason without changing the rank try this.
SELECT rank_id,
ID_code,
CAST(CONCAT(ASCII(LEFT(id_code,1)),ASCII(RIGHT(id_code,1))) AS INT) AS numeric_id_code
FROM tbl_Keys
SQL Server 2005.
I have 2 lists:
('a', 'b', 'c', 'd')
('d', 'e', 'f', 'g')
I need to make a WHERE clause with these 2 inside a dynamic SQL string, something like:
select *
from tbl
where ... afewconditionshere ...
AND anyitemfrom[('a', 'b', 'c', 'd')] is inside [('d', 'e', 'f', 'g')]
it looks kinda weird, but I get this data from a 3rd party, cannot make too big changes
the lists would always have < 20 items inside
UPDATE
a,b,c,d,e,f,g are like security things not related to the table in any way, the idea is that if you have this anyitemfrom[('a', 'b', 'c', 'd')] is inside [('d', 'e', 'f', 'g')] then you are able to view the records returned, otherwise no records should be returned
yes this condition should basically return true or false
Based on your question, it seems if you perform one IN (), then a second IN() on the filtered data set, you should get what you're looking for:
select * from
(
select * from tbl
where ... afewconditionswhere ...
and anyitemfrom in('a', 'b', 'c', 'd')
) as derived
where derived.anyitemfrom in ('d', 'e', 'f', 'g')
declare #cmd varchar(2000), #list1 varchar(100), #list2 varchar(100)
select #list1 = "('a', 'b', 'c', 'd')", #list2 = "('d', 'e', 'f', 'g')"
select #cmd =
'select * from
(
select * from tbl
where ... afewconditionswhere ...
and anyitemfrom in ' + #list1 + '
) as derived
where derived.anyitemfrom in ' + #list2
exec(#cmd)
i figured it out
i used something like :
CREATE TABLE #t (UserName VARCHAR(50))
DECLARE #sql VARCHAR(MAX)
SELECT #sql = 'INSERT INTO #t SELECT ''' + REPLACE(#UserList, ',', ''' UNION SELECT ''') + ''''
PRINT (#sql)
EXEC (#sql)
SELECT * FROM #t
IF OBJECT_ID('tempdb..#t') IS NOT NULL BEGIN DROP TABLE #t END
for both lists
then I was able to join them, and use the join's recordcount
WHERE (
SELECT COUNT(*)
FROM ( select 'a' union all select 'b' union all select 'c' union all select 'd'
intersect
select 'd' union all select 'e' union all select 'f' union all select 'g'
)
) > 0
How to loop through a select statement results to have a formatted text?
for example the select is like:
select name from table
and we want a variable #names like this:
"name1,name2,name3"
Database is SQL Server 2005
If table contains several records, i.e.:
1, name1, ..
2, name2, ..
3, name3, ..
then this query:
DECLARE #names VARCHAR(MAX)
SELECT #names = COALESCE(#names + ', ', '') + name
FROM table
will produce next result:
name1, name2, name3
See COALESCE on MSDN
This would need to be done within a function. There's no quick way to do this. Normally you would do this within your client side code.
If you plan on making a function that you do on each row form another query it will be really slow, because the function needs to be called for each row. You should work with sets of data at one time, it does all rows at one time. This is an example of how to concatenate multiple values for each row of another query:
set nocount on;
declare #t table (id int, name varchar(20), x char(1))
insert into #t (id, name, x)
select 1,'test1', 'a' union
select 1,'test1', 'b' union
select 1,'test1', 'c' union
select 2,'test2', 'a' union
select 2,'test2', 'c' union
select 3,'test3', 'b' union
select 3,'test3', 'c'
SELECT p1.id, p1.name,
stuff(
(SELECT
', ' + x
FROM #t p2
WHERE p2.id=p1.id
ORDER BY name, x
FOR XML PATH('')
)
,1,2, ''
) AS p3
FROM #t p1
GROUP BY
id, name
OUTPUT:
id name p3
----------- -------------------- -----------
1 test1 a, b, c
2 test2 a, c
3 test3 b, c