Script to get all numbers between Characters - sql

Question

I need to find a way to get the numbers between the dashes. This is the only way I know to do it, but I know that not all of our accounts are the same length. So I am just looking for a way to get everything between before, after and between the dashes. This is an example of the types of accounts we have. '2-0-200-325-0' and '1-0-1105-1500-1520' The non-digit characters are only dashes and nothing else.
declare #Department Int
declare #Account Int
declare #Company Int
declare #Location Int
declare #SubAccount Int
declare #AccountNo varchar(24) = '2-0-200-325-0'
declare #CommaPos Int
select #CommaPos = charindex('-',#accountno)
set #Company = substring(#accountno,1,#CommaPos-1)
select #Company as Company
set #Location = Substring(#AccountNo, #CommaPos+1, 1)
select #Location as Location
set #Department = Substring(#AccountNo, #CommaPos+3, 4)
select #Department as Department
set #Account = Substring(#AccountNo, #CommaPos+8, 4)
select #Account as Account
set #SubAccount = Substring(#AccountNo, #CommaPos+13, 4)
select #SubAccount as SubAccount

Answer 1

One option uses a recursive query for parsing. This properly handles the variable lenght of each part - and can easily be extended to handle more parts if needed.
-- declare the variables
declare #AccountNo varchar(24) = '2-0-200-325-0';
declare #Department Int;
declare #Account Int;
declare #Company Int;
declare #Location Int;
declare #SubAccount Int;
-- parse and assign values to variables
with cte as (
select
substring(#AccountNo + '-', 1, charindex('-', #AccountNo + '-') - 1) val,
substring(#AccountNo + '-', charindex('-', #AccountNo + '-') + 1, len(#AccountNo)) rest,
1 lvl
union all
select
substring(rest, 1, charindex('-', rest) - 1),
substring(rest, charindex('-', rest) + 1, len(rest)),
lvl + 1
from cte
where charindex('-', rest) > 0
)
select
#Company = max(case when lvl = 1 then val end),
#Location = max(case when lvl = 2 then val end),
#Department = max(case when lvl = 3 then val end),
#Account = max(case when lvl = 4 then val end),
#SubAccount = max(case when lvl = 5 then val end)
from cte;
-- check the results
select
#AccountNo AccountNo,
#Company Company,
#Location Location,
#Department Department,
#Account Account,
#SubAccount SubAccount
;
Demo on DB Fiddle:
AccountNo | Company | Location | Department | Account | SubAccount
:------------ | ------: | -------: | ---------: | ------: | ---------:
2-0-200-325-0 | 2 | 0 | 200 | 325 | 0

Answer 2

This was my approach:
--first I use a declared table variable to simulate your issue:
DECLARE #tbl TABLE(ID INT IDENTITY,ACCOUNT_NO VARCHAR(24))
INSERT INTO #tbl VALUES('2-0-200-325-0');
--The query
SELECT t.ID
,t.ACCOUNT_NO
,casted.value('x[1]','int') AS Company
,casted.value('x[2]','int') AS Location
,casted.value('x[3]','int') AS Department
,casted.value('x[4]','int') AS Account
,casted.value('x[5]','int') AS SubAccount
FROM #tbl t
CROSS APPLY(VALUES(CAST('<x>' + REPLACE(t.ACCOUNT_NO,'-','</x><x>') + '</x>' AS XML))) A(casted);
The idea in short:
We use simple string operations to transform your dash-separated list of numbers into XML.
Now we can use XQuery to retrieve each element by its position (typesafe!).
Find details here. In this link there is also a faster approach using JSON support (needs v2016+):
SELECT t.ID
,t.ACCOUNT_NO
,A.*
FROM #tbl t
CROSS APPLY OPENJSON(CONCAT('[[',REPLACE(t.ACCOUNT_NO,'-',','),']]'))
WITH(Company INT '$[0]'
,Location INT '$[1]'
,Department INT '$[2]'
,Account INT '$[3]'
,SubAccount INT '$[4]') A;
The idea of this JSON approach:
Agains we use some simple string operations to transform your string into a JSON array.
Using two array brackets ([[) allows to use OPENJSON() with a WITH clause.
The WITH clause allows to grab each fragment by its (zero-based) position (typesafe).
The WITH clause is some kind of implicit pivoting.

Answer 3

/*
-- First Create this function. This is what you need.
-- It will split a sentence into words, given a defined separator
CREATE FUNCTION [dbo].[udf_SplitString] (#Sentence varchar(max), #Separator char(1))
RETURNS #WordList TABLE (Word varchar(50))
AS
BEGIN
SET #Separator = ISNULL(#Separator, ' ')
DECLARE #Word varchar(50)
SET #Sentence = LTRIM(#Sentence) + #Separator
WHILE (CHARINDEX(#Separator, #Sentence) > 0)
BEGIN
SET #Word = SUBSTRING(#Sentence, 1, CHARINDEX(#Separator, #Sentence) - 1)
INSERT INTO #WordList SELECT LTRIM(#Word)
-- Remove word added to the List from the sentence.
SET #Sentence = SUBSTRING(#Sentence, CHARINDEX(#Separator, #Sentence) + 1, LEN(#Sentence))
SET #Sentence = LTRIM(#Sentence)
END
RETURN
END
GO
*/
DECLARE #AccountList TABLE (AccountNo varchar(20), Variable varchar(20))
INSERT INTO #AccountList VALUES
('1-0-1105-1200-1290','')
, ('1-0-1105-1500-1520','')
, ('1-0-1105-1500-1620','')
, ('1-0-1106-1200-1250','')
, ('1-0-1106-1200-1290','')
, ('1-0-1106-1500-1520','')
;
DECLARE #VariableList TABLE (OrderNo int, VariableName varchar(20))
INSERT INTO #VariableList VALUES
(1, 'Company ')
, (2, 'Location ')
, (3, 'Department ')
, (4, 'Account ')
, (5, 'SubAccount ')
;
SELECT
AccountNo
, Variable = (SELECT VariableName FROM #VariableList WHERE RowNo = OrderNo)
, Value = Word
FROM (
SELECT
RowNo = ROW_NUMBER() OVER(PARTITION BY AccountNo ORDER BY AccountNo)
, AccountNo = L.AccountNo
, Variable = ''
, Word = W.Word
FROM #AccountList L
CROSS APPLY dbo.udf_SplitString(L.AccountNo, '-') W -- Here how to use the function
) R