While going through the SQL CTEs I came across this: CODE Project CTE
The code is:
WITH ShowMessage(STATEMENT, LENGTH)
AS
(
SELECT STATEMENT = CAST('I Like ' AS VARCHAR(300)), LEN('I Like ')
UNION ALL
SELECT
CAST(STATEMENT + 'CodeProject! ' AS VARCHAR(300))
, LEN(STATEMENT) FROM ShowMessage
WHERE LENGTH < 300
)
SELECT STATEMENT, LENGTH FROM ShowMessage
or even a small modified one:
WITH ShowMessage(STATEMENT, LENGTH)
AS
(
SELECT STATEMENT = 1, LEN('I Like ')
UNION ALL
SELECT
STATEMENT + 1
, LEN(STATEMENT) FROM ShowMessage
WHERE STATEMENT < 50
)
SELECT STATEMENT, LENGTH FROM ShowMessage
The above code works perfect, when I try the code as:
with k (TT,LL)
as
(
select TT= 1, 1
union all
select TT+1,1
WHERE TT < 50
)
select TT,LL from k
My code does not work, error is that COLUMN TT does not exists. After a careful observation found that the STATEMENT is a keyword (UI showed in blue color); then I started searching online for the meaning of this keyword but could not find one (Google always throws only SELECT statement - not the STATEMENT)
Could you please explain what is this STATEMENT keyword and where/how to use it. Or please point me to the right source to learn it.
Try below query :
;WITH k (TT,LL)
as
(
SELECT 1, 1
UNION ALL
SELECT TT+1,1
FROM k --- you miss that table
WHERE TT < 50
)
SELECT TT,LL FROM k
You missed to add table i.e k.
with k (TT,LL)
as
(
select TT=1 , 1
union all
select TT+1,1 From K
WHERE TT < 50
)
select TT,LL from k
Related
I am running this SQL code on SQL Server to print alphabets from A to Z:
;with alphaCte as
(
select 'A' as letter
union all
select char(ascii(letter)+1)
from alphaCte
where letter < 'Z'
)
select * from alphaCte
I get this error:
Types don't match between the anchor and the recursive part in column "letter" of recursive query "alphaCte".
To rectify it, I have to make below change.
;with alphaCte as
(
select char(ascii('A')) as letter
union all
select char(ascii(letter)+1)
from alphaCte
where letter < 'Z'
)
select * from alphaCte
which works fine.
Could anyone please explain why my original code is throwing this datatype mismatch error?
To expand on my comment
Select column_ordinal
,name
,system_type_name
From sys.dm_exec_describe_first_result_set('select ''A'' as letter,char(ascii(''A'')+1) as letter2',null,null )
Results
column_ordinal name system_type_name
1 letter varchar(1)
2 letter2 char(1)
EDIT: Just an aside... Recursive CTEs are great but datasets are better :)
Select Top 26 C=char(64+Row_Number() Over (Order By (Select NULL)) )
From master..spt_values n
There is a column name from which I want to use to make a new column.
example:
name
asd_abceur1mz_a
asd_fxasdrasdusd3mz_a
asd_abceur10yz_a
asd_fxasdrasdusd15yz_a
The length of the column is not fixed so I assumed i have to use charindex to have a reference point from which I could trim.
What i want: at the end there is always z_a, and i need to place in a separate column the left part from z_a like this:
nameNew
eur1m
usd3m
eur10y
usd15y
The problem is that the number (in this example 1, 3, 10, 15) has 1 or two digits. I need to extract the information from name to nameNew.
After that i was thinking to make it easier to read and to output it like this:
eur_1m
usd_3m
eur_10y
usd_15y
I tried using a combination of substring and charindex, but so far without success.
SELECT *
, SUBSTRING(name, 1, ( CHARINDEX('z_a', NAME) - 1 )) AS nameNew
FROM myTable
This is for the first step, trimming the string, for the 2nd step (making it easier to read) I don't know how to target the digit and place an _.
Any help would be appreciated. Using sql server 2012
edit:
First of all thank you for your time and solutions. But your queries more or less even if they are working for 1 or 2 digits have the same problem. Consider this situation:
name
ab_dertEUR03EUR10YZ_A
if eur is two times in the string, then how can I eliminate this? Sorry for not includding this in my original post but i forgot that situation is possible and now that's a problem.
edit:
test your queries here, on this example:
http://www.sqlfiddle.com/#!3/21610/1
Please note that at the end it can be any combination of 1 or 2 digits and the letter y or m.
Ex: ab_rtgtEUR03EUR2YZ_A , ab_rtgtEUR03EUR2mZ_A, ab_rtgtEUR03EUR20YZ_A, ab_rtgtEUR03EUR20mZ_A
Some values for testing:
('ex_CHFCHF01CHF10YZ_A'), ('ab_rtgtEUR03EUR2YZ_A'), ('RON_asdRON2MZ_A'),
('tg_USDUSD04USD5YZ_A');
My understanding of your queries is that they perform something simillar to this (or at least they should)
ex_CHFCHF01CHF10YZ_A -> ex_CHFCHF01CHF10Y -> Y01FHC10FHCFHC -> Y01FHC -> CHF01Y -> CHF_01Y
RON_asdRON2MZ_A -> RON_asdRON2M -> M2NORdsa_ron -> M2NOR -> RON2M -> RON_2M
This works for one or two digits:
stuff(case
when name like '%[0-9][0-9]_z[_]a'
then left(right(name, 9), 6)
when name like '%[0-9]_z[_]a'
then left(right(name, 8), 5)
end, 4, 0, '_')
You can use a combination of substring , reverse and charindex.
SQL Fiddle
select substring(namenew,1,3) + '_' + substring(namenew, 4, len(namenew))
from (
select
case when name like '%[0-9][0-9]_z[_]a' then
reverse(substring(reverse(name), charindex('a_z',reverse(name)) + 3, 6))
when name like '%[0-9]_z[_]a' then
reverse(substring(reverse(name), charindex('a_z',reverse(name)) + 3, 5))
end as namenew
from myTable
) t
Try it like this:
declare #tbl TABLE(name VARCHAR(100));
insert into #tbl VALUES
('asd_abceur1mz_a')
,('asd_fxasdrasdusd3mz_a')
,('asd_abceur10yz_a')
,('asd_fxasdrasdusd15yz_a')
,('ab_dertEUR03EUR10YZ_A');
WITH CutOfThreeAtTheEnd AS
(
SELECT LEFT(name,LEN(name)-3) AS nameNew
FROM #tbl
)
,Max6CharsFromEnd AS
(
SELECT RIGHT(nameNew,6) AS nameNew
FROM CutOfThreeAtTheEnd
)
SELECT nameNew
,FirstNumber.Position
,Parts.*
,Parts.FrontPart + '_' + Parts.BackPart AS FinalString
FROM Max6CharsFromEnd
CROSS APPLY
(
SELECT MIN(x)
FROM
(
SELECT CHARINDEX('0',nameNew,1) AS x
UNION SELECT CHARINDEX('1',nameNew,1)
UNION SELECT CHARINDEX('2',nameNew,1)
UNION SELECT CHARINDEX('3',nameNew,1)
UNION SELECT CHARINDEX('4',nameNew,1)
UNION SELECT CHARINDEX('5',nameNew,1)
UNION SELECT CHARINDEX('6',nameNew,1)
UNION SELECT CHARINDEX('7',nameNew,1)
UNION SELECT CHARINDEX('8',nameNew,1)
UNION SELECT CHARINDEX('9',nameNew,1)
) AS tbl
WHERE x>0
) AS FirstNumber(Position)
CROSS APPLY(SELECT SUBSTRING(nameNew,FirstNumber.Position,1000) AS BackPart
,SUBSTRING(nameNew,FirstNumber.Position-3,3) AS FrontPart) AS Parts
this is the result:
nameNew Position BackPart FrontPart FinalString
ceur1m 5 1m eur eur_1m
dusd3m 5 3m usd usd_3m
eur10y 4 10y eur eur_10y
usd15y 4 15y usd usd_15y
EUR10Y 4 10Y EUR EUR_10Y
I have a table EmployeeTable.
If I want only that records where employeename have character of 1 to 5
will be palindrome and there also condition like total character is more then 10 then 4 to 8 if character less then 7 then 2 to 5 and if character less then 5 then all char will be checked and there that are palindrome then only display.
Examples :- neen will be display
neetan not selected
kiratitamara will be selected
I try this something on string function like FOR first case like name less then 5 character long
SELECT SUBSTRING(EmployeeName,1,5),* from EmaployeeTable where
REVERSE (SUBSTRING(EmployeeName,1,5))=SUBSTRING(EmployeeName,1,5)
I want to do that without string functions,
Can anyone help me on this?
You need at least SUBSTRING(), I have a solution like this:
(In SQL Server)
DECLARE #txt varchar(max) = 'abcba'
;WITH CTE (cNo, cChar) AS (
SELECT 1, SUBSTRING(#txt, 1, 1)
UNION ALL
SELECT cNo + 1, SUBSTRING(#txt, cNo + 1, 1)
FROM CTE
WHERE SUBSTRING(#txt, cNo + 1, 1) <> ''
)
SELECT COUNT(*)
FROM (
SELECT *, ROW_NUMBER() OVER (ORDER BY cNo DESC) as cRevNo
FROM CTE t1 CROSS JOIN
(SELECT Max(cNo) AS strLength FROM CTE) t2) dt
WHERE
dt.cNo <= dt.strLength / 2
AND
dt.cChar <> (SELECT dti.cChar FROM CTE dti WHERE dti.cNo = cRevNo)
The result will shows the count of differences and 0 means no differences.
Note :
Current solution is Non-Case-Sensitive for change it to a Case-Sensitive you need to check the strings in a case-sensitive collation like Latin1_General_BIN
You can use this solution as a SVF or something like that.
I dont realy understand why you dont want to use string functions in your query, but here is one solution. Compute everything beforehand:
Add Column:
ALTER TABLE EmployeeTable
ADD SubString AS
SUBSTRING(EmployeeName,
(
CASE WHEN LEN(EmployeeName)>10
THEN 4
WHEN LEN(EmployeeName)>7
THEN 2
ELSE 1 END
)
,
(
CASE WHEN LEN(EmployeeName)>10
THEN 8
WHEN LEN(EmployeeName)>7
THEN 5
ELSE 5 END
)
PERSISTED
GO
ALTER TABLE EmployeeTable
ADD Palindrome AS
REVERSE(SUBSTRING(EmployeeName,
(
CASE WHEN LEN(EmployeeName)>10
THEN 4
WHEN LEN(EmployeeName)>7
THEN 2
ELSE 1 END
)
,
(
CASE WHEN LEN(EmployeeName)>10
THEN 8
WHEN LEN(EmployeeName)>7
THEN 5
ELSE 5 END
)) PERSISTED
GO
Then your query will looks like:
SELECT * from EmaployeeTable
where Palindrome = SubString
BUT!
This is not a good idea. Please tell us, why you dont want to use string functios.
You could do it building a list of palindrome words using a recursive query that generates palindrome words till a length o n characters and then selects employees with the name matching a palindrome word. This may be a really inefficient way, but it does the trick
This is a sample query for Oracle, PostgreSQL should support this feature as well with little differences on syntax. I don't know about other RDBMS.
with EmployeeTable AS (
SELECT 'ADA' AS employeename
FROM DUAL
UNION ALL
SELECT 'IDA' AS employeename
FROM DUAL
UNION ALL
SELECT 'JACK' AS employeename
FROM DUAL
), letters as (
select chr(ascii('A') + rownum - 1) as letter
from dual
connect by ascii('A') + rownum - 1 <= ascii('Z')
), palindromes(word, len ) as (
SELECT WORD, LEN
FROM (
select CAST(NULL AS VARCHAR2(100)) as word, 0 as len
from DUAL
union all
select letter as word, 1 as len
from letters
)
union all
select l.letter||p.word||l.letter AS WORD, len + 1 AS LEN
from palindromes p
cross join letters l
where len <= 4
)
SEARCH BREADTH FIRST BY word SET order1
CYCLE word SET is_cycle TO 'Y' DEFAULT 'N'
select *
from EmployeeTable
WHERE employeename IN (
SELECT WORD
FROM palindromes
)
DECLARE #cPalindrome VARCHAR(100) = 'SUBI NO ONIBUS'
SET #cPalindrome = REPLACE(#cPalindrome, ' ', '')
;WITH tPalindromo (iNo) AS (
SELECT 1
WHERE SUBSTRING(#cPalindrome, 1, 1) = SUBSTRING(#cPalindrome, LEN(#cPalindrome), 1)
UNION ALL
SELECT iNo + 1
FROM tPalindromo
WHERE SUBSTRING(#cPalindrome, iNo + 1, 1) = SUBSTRING(#cPalindrome, LEN(#cPalindrome) - iNo, 1)
AND LEN(#cPalindrome) > iNo
)
SELECT IIF(MAX(iNo) = LEN(#cPalindrome), 'PALINDROME', 'NOT PALINDROME')
FROM tPalindromo
Is there an elegant way in SQL Server to find all the distinct characters in a single varchar(50) column, across all rows?
Bonus points if it can be done without cursors :)
For example, say my data contains 3 rows:
productname
-----------
product1
widget2
nicknack3
The distinct inventory of characters would be "productwigenka123"
Here's a query that returns each character as a separate row, along with the number of occurrences. Assuming your table is called 'Products'
WITH ProductChars(aChar, remain) AS (
SELECT LEFT(productName,1), RIGHT(productName, LEN(productName)-1)
FROM Products WHERE LEN(productName)>0
UNION ALL
SELECT LEFT(remain,1), RIGHT(remain, LEN(remain)-1) FROM ProductChars
WHERE LEN(remain)>0
)
SELECT aChar, COUNT(*) FROM ProductChars
GROUP BY aChar
To combine them all to a single row, (as stated in the question), change the final SELECT to
SELECT aChar AS [text()] FROM
(SELECT DISTINCT aChar FROM ProductChars) base
FOR XML PATH('')
The above uses a nice hack I found here, which emulates the GROUP_CONCAT from MySQL.
The first level of recursion is unrolled so that the query doesn't return empty strings in the output.
Use this (shall work on any CTE-capable RDBMS):
select x.v into prod from (values('product1'),('widget2'),('nicknack3')) as x(v);
Test Query:
with a as
(
select v, '' as x, 0 as n from prod
union all
select v, substring(v,n+1,1) as x, n+1 as n from a where n < len(v)
)
select v, x, n from a -- where n > 0
order by v, n
option (maxrecursion 0)
Final Query:
with a as
(
select v, '' as x, 0 as n from prod
union all
select v, substring(v,n+1,1) as x, n+1 as n from a where n < len(v)
)
select distinct x from a where n > 0
order by x
option (maxrecursion 0)
Oracle version:
with a(v,x,n) as
(
select v, '' as x, 0 as n from prod
union all
select v, substr(v,n+1,1) as x, n+1 as n from a where n < length(v)
)
select distinct x from a where n > 0
Given that your column is varchar, it means it can only store characters from codes 0 to 255, on whatever code page you have. If you only use the 32-128 ASCII code range, then you can simply see if you have any of the characters 32-128, one by one. The following query does that, looking in sys.objects.name:
with cteDigits as (
select 0 as Number
union all select 1 as Number
union all select 2 as Number
union all select 3 as Number
union all select 4 as Number
union all select 5 as Number
union all select 6 as Number
union all select 7 as Number
union all select 8 as Number
union all select 9 as Number)
, cteNumbers as (
select U.Number + T.Number*10 + H.Number*100 as Number
from cteDigits U
cross join cteDigits T
cross join cteDigits H)
, cteChars as (
select CHAR(Number) as Char
from cteNumbers
where Number between 32 and 128)
select cteChars.Char as [*]
from cteChars
cross apply (
select top(1) *
from sys.objects
where CHARINDEX(cteChars.Char, name, 0) > 0) as o
for xml path('');
If you have a Numbers or Tally table which contains a sequential list of integers you can do something like:
Select Distinct '' + Substring(Products.ProductName, N.Value, 1)
From dbo.Numbers As N
Cross Join dbo.Products
Where N.Value <= Len(Products.ProductName)
For Xml Path('')
If you are using SQL Server 2005 and beyond, you can generate your Numbers table on the fly using a CTE:
With Numbers As
(
Select Row_Number() Over ( Order By c1.object_id ) As Value
From sys.columns As c1
Cross Join sys.columns As c2
)
Select Distinct '' + Substring(Products.ProductName, N.Value, 1)
From Numbers As N
Cross Join dbo.Products
Where N.Value <= Len(Products.ProductName)
For Xml Path('')
Building on mdma's answer, this version gives you a single string, but decodes some of the changes that FOR XML will make, like & -> &.
WITH ProductChars(aChar, remain) AS (
SELECT LEFT(productName,1), RIGHT(productName, LEN(productName)-1)
FROM Products WHERE LEN(productName)>0
UNION ALL
SELECT LEFT(remain,1), RIGHT(remain, LEN(remain)-1) FROM ProductChars
WHERE LEN(remain)>0
)
SELECT STUFF((
SELECT N'' + aChar AS [text()]
FROM (SELECT DISTINCT aChar FROM Chars) base
ORDER BY aChar
FOR XML PATH, TYPE).value(N'.[1]', N'nvarchar(max)'),1, 1, N'')
-- Allow for a lot of recursion. Set to 0 for infinite recursion
OPTION (MAXRECURSION 365)
I have a coworker looking for this, and I don't recall ever running into anything like that.
Is there a reasonable technique that would let you simulate it?
SELECT PRODUCT(X)
FROM
(
SELECT 3 X FROM DUAL
UNION ALL
SELECT 5 X FROM DUAL
UNION ALL
SELECT 2 X FROM DUAL
)
would yield 30
select exp(sum(ln(col)))
from table;
edit:
if col always > 0
DECLARE #a int
SET #a = 1
-- re-assign #a for each row in the result
-- as what #a was before * the value in the row
SELECT #a = #a * amount
FROM theTable
There's a way to do string concat that is similiar:
DECLARE #b varchar(max)
SET #b = ""
SELECT #b = #b + CustomerName
FROM Customers
Here's another way to do it. This is definitely the longer way to do it but it was part of a fun project.
You've got to reach back to school for this one, lol. They key to remember here is that LOG is the inverse of Exponent.
LOG10(X*Y) = LOG10(X) + LOG10(Y)
or
ln(X*Y) = ln(X) + ln(Y) (ln = natural log, or simply Log base 10)
Example
If X=5 and Y=6
X * Y = 30
ln(5) + ln(6) = 3.4
ln(30) = 3.4
e^3.4 = 30, so does 5 x 6
EXP(3.4) = 30
So above, if 5 and 6 each occupied a row in the table, we take the natural log of each value, sum up the rows, then take the exponent of the sum to get 30.
Below is the code in a SQL statement for SQL Server. Some editing is likely required to make it run on Oracle. Hopefully it's not a big difference but I suspect at least the CASE statement isn't the same on Oracle. You'll notice some extra stuff in there to test if the sign of the row is negative.
CREATE TABLE DUAL (VAL INT NOT NULL)
INSERT DUAL VALUES (3)
INSERT DUAL VALUES (5)
INSERT DUAL VALUES (2)
SELECT
CASE SUM(CASE WHEN SIGN(VAL) = -1 THEN 1 ELSE 0 END) % 2
WHEN 1 THEN -1
ELSE 1
END
* CASE
WHEN SUM(VAL) = 0 THEN 0
WHEN SUM(VAL) IS NOT NULL THEN EXP(SUM(LOG(ABS(CASE WHEN SIGN(VAL) <> 0 THEN VAL END))))
ELSE NULL
END
* CASE MIN(ABS(VAL)) WHEN 0 THEN 0 ELSE 1 END
AS PRODUCT
FROM DUAL
The accepted answer by tuinstoel is correct, of course:
select exp(sum(ln(col)))
from table;
But notice that if col is of type NUMBER, you will find tremendous performance improvement when using BINARY_DOUBLE instead. Ideally, you would have a BINARY_DOUBLE column in your table, but if that's not possible, you can still cast col to BINARY_DOUBLE. I got a 100x improvement in a simple test that I documented here, for this cast:
select exp(sum(ln(cast(col as binary_double))))
from table;
Is there a reasonable technique that would let you simulate it?
One technique could be using LISTAGG to generate product_expression string and XMLTABLE + GETXMLTYPE to evaluate it:
WITH cte AS (
SELECT grp, LISTAGG(l, '*') AS product_expression
FROM t
GROUP BY grp
)
SELECT c.*, s.val AS product_value
FROM cte c
CROSS APPLY(
SELECT *
FROM XMLTABLE('/ROWSET/ROW/*'
PASSING dbms_xmlgen.getXMLType('SELECT ' || c.product_expression || ' FROM dual')
COLUMNS val NUMBER PATH '.')
) s;
db<>fiddle demo
Output:
+------+---------------------+---------------+
| GRP | PRODUCT_EXPRESSION | PRODUCT_VALUE |
+------+---------------------+---------------+
| b | 2*6 | 12 |
| a | 3*5*7 | 105 |
+------+---------------------+---------------+
More roboust version with handling single NULL value in the group:
WITH cte AS (
SELECT grp, LISTAGG(l, '*') AS product_expression
FROM t
GROUP BY grp
)
SELECT c.*, s.val AS product_value
FROM cte c
OUTER APPLY(
SELECT *
FROM XMLTABLE('/ROWSET/ROW/*'
passing dbms_xmlgen.getXMLType('SELECT ' || c.product_expression || ' FROM dual')
COLUMNS val NUMBER PATH '.')
WHERE c.product_expression IS NOT NULL
) s;
db<>fiddle demo
*CROSS/OUTER APPLY(Oracle 12c) is used for convenience and could be replaced with nested subqueries.
This approach could be used for generating different aggregation functions.
There are many different implmentations of "SQL". When you say "does sql have" are you referring to a specific ANSI version of SQL, or a vendor specific implementation. DavidB's answer is one that works in a few different environments I have tested but depending on your environment you could write or find a function exactly like what you are asking for. Say you were using Microsoft SQL Server 2005, then a possible solution would be to write a custom aggregator in .net code named PRODUCT which would allow your original query to work exactly as you have written it.
In c# you might have to do:
SELECT EXP(SUM(LOG([col])))
FROM table;