December 2, 2008 at 9:43 pm
Comments posted to this topic are about the item Increment a string
December 3, 2008 at 7:53 am
How do you plan on handling when you need to increment 'ZZ'?
Toni
December 5, 2008 at 10:10 am
Here is a stab at this issue as a function. If you pad the input with your zero symbol, you will get padded output, but it will grow the "number" beyond the bounds of the padding up to the length defined for the @Ticket input. Thus, 'ZZ' in returns '100' while '00ZZ' in returns '0100'.
CREATE FUNCTION dbo.NextTicket( @Ticket VARCHAR(128) ) RETURNS VARCHAR(128) AS
BEGIN
DECLARE @digits TABLE( value INT IDENTITY(0,1) NOT NULL PRIMARY KEY, symbol CHAR(1) NOT NULL UNIQUE );
DECLARE @place INT, @zero_digit CHAR(1), @highest_digit CHAR(1);
-- 1. build a table of symbols that compose the "digits" of our ticket numbers
-- Note that you could impose artifical order, it just happens that sorting by
-- the value gives me the order I want.
INSERT INTO @digits ( symbol )
SELECT ch
FROM (
SELECT '0' AS ch UNION SELECT '1' 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'
) AS symbols
ORDER BY ch ASC;
SELECT @zero_digit = symbol FROM @digits WHERE value = 0;
SELECT TOP 1 @highest_digit = symbol FROM @digits ORDER BY value DESC;
-- 2. assume valid input (functions can't throw errors) -----------------------
SET @Ticket = RTRIM(LTRIM(@Ticket));
-- 3. carry the 1 -------------------------------------------------------------
SET @place = LEN(@Ticket);
WHILE @place > 0 AND SUBSTRING(@Ticket, @place, 1) = @highest_digit
BEGIN
SET @Ticket = STUFF(@Ticket, @place, 1, @zero_digit);
SET @place = @place -1;
END
-- 4. final increment ---------------------------------------------------------
IF @place = 0 BEGIN
DECLARE @one_digit CHAR(1), @length INT;
SELECT @one_digit = symbol FROM @digits WHERE value = 1;
SET @length = LEN(@Ticket);
SET @Ticket = @one_digit + @Ticket; -- if this overflows, it will truncate the right most "digit"
IF @length = LEN(@Ticket) -- still the same length only if overflowed...
SET @Ticket = REPLICATE(@zero_digit, @length); -- roll over back to zero
END ELSE BEGIN
DECLARE @next_digit CHAR(1);
SET @next_digit = (
SELECT symbol FROM @digits WHERE value = (
SELECT value +1 FROM @digits WHERE symbol = SUBSTRING(@Ticket, @place, 1)));
SET @Ticket = STUFF(@Ticket, @place, 1, @next_digit);
END
-- 5. Return the new Ticket number
RETURN @Ticket;
END
December 6, 2008 at 8:51 am
Hi Dennis. Interesting way to carry the digits. I had done another one that had a fixed length for the string to be increment (e.g. 2 chars). If overflow occurred, it returns '**' (the pseudo-standard overflow chars) or if any invalid characters (not in the set of incrementable characters) were passed, a
'--' is returned.
The function I built used strings and indexes into them versus creating and working through a table. In a simple comparison from using the below statements while executing the code within your and my functions, not using a table yields much better performance (particularly if the function were invoked with any regularity).
DBCC DROPCLEANBUFFERS
GO
SET STATISTICS IO ON
SET STATISTICS TIME ON
-----------------------------------------------------------------------------------------
RESULTS Elapsed | Statements Run | Logical Read | Scans | CPU
-----------------------------------------------------------------------------------------
Using table 126 ms | 23 | 118 | 4 | 48
Using string 0 ms | 11 | 0 | 0 | 0
The function I created is here:
/* Remove any prior version of function */
IF EXISTS (SELECT * FROM dbo.sysobjects WHERE id = OBJECT_ID(N'[dbo].[fn_NextSerial]')
AND xtype IN (N'FN', N'IF', N'TF'))
DROP FUNCTION [dbo].[fn_NextSerial]
GO
CREATE FUNCTION fn_NextSerial (@currentboth CHAR(2))
RETURNS char(2)
AS
/* Given the current value of a two-character string (@currentboth),
return the next higher value based on the string of valid chars (@alphas)
Example: fn_NextSerial('AZ') will return 'BA'
Note: If the string can not be incremented any further, '**' is returned
If the string contains invalid characters, '--' is returned
*/
BEGIN
DECLARE @alphas CHAR(36) -- valid alpha values (length must match that of string assigned below)
SET @alphas = '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ' -- valid chars for incrementing string
DECLARE @current1 INT -- index in valid alpha values of 1st char in string
DECLARE @current2 INT -- index in valid alpha values of 2nd char in string
SET @current1 = CHARINDEX(SUBSTRING(@currentboth,1,1),@alphas)
SET @current2 = CHARINDEX(SUBSTRING(@currentboth,2,1),@alphas)
/* If any invalid or null input, return '--' */
IF @current1 = 0 OR @current2 = 0 OR @currentboth IS NULL
SET @currentboth = '--'
ELSE
BEGIN
IF @current2 = LEN(@alphas) -- Is the 2nd char at the last valid string position?
BEGIN
IF @current1 <> LEN(@alphas) -- Can the 1st char still be incremented?
BEGIN
/* Increment the 1st char position and set the 2nd char to starting char */
SET @currentboth = SUBSTRING(@alphas,(@current1+1),1)+
SUBSTRING(@alphas,1,1)
END
ELSE
SET @currentboth = '**' -- return '**' to show no more incrementing possible
END
ELSE
/* Leave 1st char unchanged and increment the 2nd char of the string */
SET @currentboth= SUBSTRING(@currentboth,1,1) + SUBSTRING(@alphas,@current2+1,1)
END
RETURN @currentboth -- return the next value or
-- '**' if already at highest value or
-- '--' if invalid chars passed
END
GO
Select 'Next after 00 is:' + dbo.fn_NextSerial('00') -- '01'
Select 'Next after 0Y is:' + dbo.fn_NextSerial('0Y') -- '0Z'
Select 'Next after 0Z is:' + dbo.fn_NextSerial('0Z') -- '10'
Select 'Next after ZY is:' + dbo.fn_NextSerial('ZY') -- 'ZZ'
Select 'Next after ZZ is:' + dbo.fn_NextSerial('ZZ') -- '**'
Select 'Next after '' Z'' is:' + dbo.fn_NextSerial(' Z') -- '--'
Select 'Next after @0 is:' + dbo.fn_NextSerial('@0') -- '--'
Select 'Next after A$ is:' + dbo.fn_NextSerial('A$') -- '--'
Select 'Next after null is:' + dbo.fn_NextSerial(NULL) -- '--'
I modified your function to use the string and it brought the usage back comparable to what my function had achieved for CPU, IO and elapsed time. The statements that were changed are commented out and the replacements are right next to them.
If you would want to adopt returning a '**' for overflow or '--' for invalid chars, then you could use code similar to the checks in mine. Here is the heart of your function (my interpretation) using the string versus a table:
--DBCC DROPCLEANBUFFERS -- clear the buffers
--go
/* Prep as if we are calling the function */
declare @Ticket varchar(128)
set statistics io on
set statistics time on
set @Ticket = 'ZZZ' -- change this and rerun to test different cases
select 'Input is ' + @Ticket
/* Invoke the statements from the body of the function */
-- DECLARE @digits TABLE( value INT IDENTITY(0,1) NOT NULL PRIMARY KEY, symbol CHAR(1) NOT NULL UNIQUE );
DECLARE @place INT, @zero_digit CHAR(1), @highest_digit CHAR(1);
-- 1. build a table of symbols that compose the "digits" of our ticket numbers
-- Note that you could impose artifical order, it just happens that sorting by
-- the value gives me the order I want.
/* INSERT INTO @digits ( symbol )
SELECT ch
FROM (
SELECT '0' AS ch UNION SELECT '1' 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'
) AS symbols
ORDER BY ch ASC;
*/
DECLARE @digits CHAR(36) -- valid alpha values (length must match that of string assigned below)
SET @digits = '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ' -- valid chars for incrementing string
-- SELECT @zero_digit = symbol FROM @digits WHERE value = 0;
SET @zero_digit = SUBSTRING(@digits,1,1)
-- SELECT TOP 1 @highest_digit = symbol FROM @digits ORDER BY value DESC;
SET @highest_digit = SUBSTRING(@digits,LEN(@digits),1)
-- 2. assume valid input (functions can't throw errors) -----------------------
SET @Ticket = RTRIM(LTRIM(@Ticket));
-- 3. carry the 1 -------------------------------------------------------------
SET @place = LEN(@Ticket);
WHILE @place > 0 AND SUBSTRING(@Ticket, @place, 1) = @highest_digit
BEGIN
SET @Ticket = STUFF(@Ticket, @place, 1, @zero_digit);
SET @place = @place -1;
END
-- 4. final increment ---------------------------------------------------------
IF @place = 0 BEGIN
DECLARE @one_digit CHAR(1), @length INT;
-- SELECT @one_digit = symbol FROM @digits WHERE value = 1;
SET @one_digit = SUBSTRING(@digits,2,1)
SET @length = LEN(@Ticket);
SET @Ticket = @one_digit + @Ticket; -- if this overflows, it will truncate the right most "digit"
IF @length = LEN(@Ticket) -- still the same length only if overflowed...
SET @Ticket = REPLICATE(@zero_digit, @length); -- roll over back to zero
END
ELSE BEGIN
DECLARE @next_digit CHAR(1);
--SET @next_digit =
-- (SELECT symbol FROM @digits WHERE value = (
-- SELECT value +1 FROM @digits WHERE symbol = SUBSTRING(@Ticket, @place, 1)));
SET @next_digit =
SUBSTRING(@digits,CHARINDEX(SUBSTRING(@Ticket,@place,1),@digits)+1,1)
SET @Ticket = STUFF(@Ticket, @place, 1, @next_digit);
END
/* Put out the result */
select 'Answer is ' + @Ticket
set statistics io off
set statistics time off
Toni
December 8, 2008 at 1:47 pm
Nice, pulling out the table is a big improvement. In looking at that, I was thinking about doing more than just adding one. So I created an addition function that allows upto 128 digit numbers in base 36, and then a separate function to constrain it to ZZ max value for Tickets.
Your solution still beats mine (4ms on my test machine verses 9ms for my solution).
Function for base36 addition
ALTER FUNCTION dbo.Base36Add( @value1 varchar(128), @value2 varchar(128) ) RETURNS varchar(128) AS
BEGIN
DECLARE @digits char(36); SET @digits = '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ';
DECLARE @sum varchar(16), @i int, @s-2 int, @d1 int, @d2 int, @r int;
-- 1. null values are invalid -------------------------------------------------
IF @value1 IS NULL OR @value2 IS NULL
GOTO Exit_Error;
-- 2. pad so they are the same length -----------------------------------------
IF len(@value1) < len(@value2)
SET @value1 = replicate( substring(@digits,1,1), len(@value2)-len(@value1) ) + @value1;
ELSE IF len(@value2) < len(@value1)
SET @value2 = replicate( substring(@digits,1,1), len(@value1)-len(@value2) ) + @value2;
-- 3. sum and carry -----------------------------------------------------------
SELECT
@sum = '', /* init string for prepending */
@i = len(@value1), /* start at least signifigant place */
@r = 0; /* start with no remainder */
WHILE 0 < @i BEGIN /* for each place */
-- a. convert the sum of the remainder plus the current place in both values to an integer
SELECT
@d1 = charindex( substring(@value1,@i,1), @digits ) - 1,
@d2 = charindex( substring(@value2,@i,1), @digits ) - 1;
IF (coalesce(@d1,-1) = -1 OR coalesce(@d2,-1) = -1)
GOTO Exit_Error;
-- b. find the resulting total for this place
SET @sum = substring(@digits, (@s % 36) + 1, 1)
+ @sum;
-- c. find the remainder to carry forward
SET @r = (@s / 36);
-- d. next place...
SET @i = @i - 1;
END
-- 4. final remainder?
IF 0 < @r BEGIN
IF len(@sum) < 128 /* defined return value size */
SET @sum = substring(@digits, @r + 1, 1) + @sum;
ELSE
SET @sum = '**'; /* overflow */
END
Exit_Success:
RETURN @sum;
Exit_Error:
RETURN '--'; /* invalid input */
END
Function to constrict Ticket numbers to 00 through ZZ (0 to 1296)
CREATE FUNCTION [dbo].[NextTicket]( @Ticket varchar(2) ) RETURNS varchar(2) AS
BEGIN
IF @Ticket = 'ZZ'
RETURN '**';
RETURN dbo.Base36Add( @Ticket, '1' );
END
Test queries for timing
SELECT
dbo.NextTicket('00') AS Expect_01,
dbo.NextTicket('0Y') AS Expect_0Z,
dbo.NextTicket('0Z') AS Expect_10,
dbo.NextTicket('ZY') AS Expect_ZZ,
dbo.NextTicket('ZZ') AS Expect_Overflow,
dbo.NextTicket(' Z') AS Expect_Invalid_01,
dbo.NextTicket('@0') AS Expect_Invalid_02,
dbo.NextTicket('A$') AS Expect_Invalid_03,
dbo.NextTicket(NULL) AS Expect_Invalid_04
SELECT
dbo.fn_NextSerial('00') AS Expect_01,
dbo.fn_NextSerial('0Y') AS Expect_0Z,
dbo.fn_NextSerial('0Z') AS Expect_10,
dbo.fn_NextSerial('ZY') AS Expect_ZZ,
dbo.fn_NextSerial('ZZ') AS Expect_Overflow,
dbo.fn_NextSerial(' Z') AS Expect_Invalid_01,
dbo.fn_NextSerial('@0') AS Expect_Invalid_02,
dbo.fn_NextSerial('A$') AS Expect_Invalid_03,
dbo.fn_NextSerial(NULL) AS Expect_Invalid_04
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply