SQL Server query to delete text from text column - sql

Question

I have a SQL Server database with a table feedback that contains a text column comment. In that column I have tag data, for example
This is my record <tag>Random characters are here</tag> with information.
How do I write a query to update all of these records to remove the <tag></tag> and all of the text in between?
I'd like to write this to a different 'temporary' table to first verify the changes and then update the original table.
I am running SQL Server 2014 Express.
Thank you

Answer 1

Here is a function to remove tags..
CREATE FUNCTION [dbo].[RemoveTag](#text NVARCHAR(MAX), #tag as nvarchar(max))
RETURNS NVARCHAR(MAX)
AS
BEGIN
declare #startTagIndex as int
declare #endTagIndex as int
set #startTagIndex = CHARINDEX('<' + #tag + '>', #text)
if(#startTagIndex > 0) BEGIN
set #endTagIndex = CHARINDEX('</' + #tag + '>', #text, #startTagIndex)
if(#endTagIndex > 0) BEGIN
return LEFT(#text, #startTagIndex - 1) + RIGHT(#text, len(#text) - len(#tag) - #endTagIndex - 2)
END
END
return #text
END
Later you can use it like:
Update table set field = dbo.RemoveTag(field, 'tag')
If you want to write fields to other table then:
CREATE TABLE dbo.OtherTable (
OtherField nvarchar(MAX) NOT NULL
)
GO
INSERT INTO OtherTable (OtherField)
SELECT dbo.RemoveTag(field, 'tag') from table

Answer 2

Making a lot assumptions about the format of your string. But if they're valid then this is very simple:
left(s, charindex('<tag>', s - 1)) +
substring(s, charindex('</tag>', s) + 6, len(s))
Obviously we're basically assuming that the search strings appear only once and in the correct order. There's also an assumption that there will be matches. Also, I used len(s) as an easy upper bound on the number of characters to take from the right. You could just hard-code something appropriate if you felt like it since SQL Server doesn't error for going past the end. s is just a stand in for your char column.
http://sqlfiddle.com/#!3/771a3/8
Not sure if extra whitespace is going to be an issue so you might want to trim and add a space character in the middle.
rtrim(left(s, charindex('<tag>', s) - 1)) + ' ' +
ltrim(substring(s, charindex('</tag>', s) + 6, len(s)))

Answer 3

You can use CHARINDEX to find where your tags start and stop, SUBSTRING to get all text between < and >, and REPLACE to swap out the substring for ''.
Select Field,
Substring(FIELD, charindex('<', Field), CHARINDEX('>', Field,
(CHARINDEX('>', FIELD)) + 1) - charindex('<', Field)+1) as ToRemove,
replace (Field, Substring(FIELD, charindex('<', Field), CHARINDEX('>',
Field, (CHARINDEX('>', FIELD)) + 1) - charindex('<', Field)+1), '')
as FinalResult
from TableName
The output will be three columns, Field, ToRemove and FinalResult, but nothing will actually be updated.
I think the only way this will fail is if you have nested tags. <b><i>sometext</i></b>
To actually make the change:
Update #TableName set Field = replace (Field, Substring(FIELD, charindex('<', Field), CHARINDEX('>', Field, (CHARINDEX('>', FIELD)) + 1) - charindex('<', Field)+1), '')
Tested on SQL Server 2012.