SQL Server: String Manipulation, Unpivoting

Question

I have a column called body, which contains body content for our CMS. The data looks like:

...{cloak:id=1.1.1}...{cloak}...{cloak:id=1.1.2}...{cloak}...{cloak:id=1.1.3}...{cloak}...

A moderately tweaked for readability example:

## h5. A formal process for approving and testing all external network connections and changes to the firewall and router configurations? 
{toggle-cloak:id=1.1.1}{tree-plus-icon} *Compliance:* {color:red}{*}Partial{*}{color} (?) 
{cloak:id=1.1.1} || Date: | 2010-03-15 || || Owner: | Brian || || Researched by: | || || Narrative: | Jira tickets are normally used to approve and track network changes\\ || || Artifacts: | Jira.bccampus.ca\\ || || Recommendation: | Need to update policy that no Jira = no change\\ || || Proposed Remedy(ies): | || || Approved Remedy(ies): | || || Date: | || || Reviewed by: | || || Remarks/comments: | || 
{cloak}## h5. Current network diagrams with all connections to cardholder data, including any wireless networks? 
{toggle-cloak:id=1.1.2}{tree-plus-icon} *Compliance:* {color:red}{*}TBD{*}{color} (?) 
{cloak:id=1.1.2}

I'd like to get the cloak values out in the following format:

 requirement_num
 -----------------
 1.1.1
 1.1.2
 1.1.3

I'm looking at using UNIONs - does anyone have a better recommendation?

Forgot to mention:

• I can't use regex, because CLR isn't enabled on the database.
• The numbers aren't sequencial. The current record jumps from 1.1.6 to 1.2.1
• I can guarantee that there will be a pair for each requirement number - {toggle-cloak:id=x.y.z} and {cloak:id=x.y.z}. I'm interested in what is between the {cloak:id=x.y.z} and {cloak} tags.

Answer 1

I'd probably use a function for this. Something like:

create function [dbo].[getCloaks]
(
@String     varchar(8000)
)
returns @tbl table (s varchar(1000))
as
begin
declare @i int, @j int, @k int
    select  @i = 1
    while charindex('{cloak:id=', @String, @i) > 0
    begin
        select @j = charindex('{cloak:id=', @String, @i)
        select @k = charindex('}', @String, @j)
        insert  @tbl select substring(@String, @j + 10, @k - @j - 10)
        select  @i = @k + 1
    end
    return
end

Answer 2

Here's a recursive CTE version that will hopefully be a little faster than an iterative UDF:

DECLARE @Data nvarchar(1000)

SET @Data = N'...{cloak:id=1.1.1}...{cloak}...{cloak:id=1.1.2}...{cloak}...' +
            N'{cloak:id=1.1.3}...{cloak}...'

;WITH Cloak_CTE AS
(
    SELECT
        CAST(NULL AS nvarchar(50)) AS requirement_num,
        CHARINDEX('{cloak:id=', @Data) AS start_index,
        CHARINDEX('}', @Data, CHARINDEX('{cloak:id=', @Data)) AS end_index

    UNION ALL

    SELECT
        CAST(SUBSTRING(@Data, start_index + 10,
            end_index - start_index - 10) AS nvarchar(50)),
        CHARINDEX('{cloak:id=', @Data, end_index + 1),
        CHARINDEX('}', @Data, CHARINDEX('{cloak:id=', @Data, end_index + 1))
    FROM Cloak_CTE
    WHERE start_index > 0
)
SELECT requirement_num
FROM Cloak_CTE
WHERE requirement_num IS NOT NULL

Should be relatively straightforward to adjust for multiple rows or different patterns, or put this into an inline UDF.