SQL LIKE statement to text field fomr CF MX 6.1

Question

Hi.
Here's my scenario. I am running SQL Server 2000 SP4 on a
Windows 2003 Server running IIS 6. Using ColdFusion MX 6.1 Updater
version.
I am storing HTML page code (some complete, some just clips)
in a text type field.
I need to check to see if someone has already saved that
exact code already before. So I do the following query:
<cfquery name="checkforexisting"
datasource="mydatasource">
SELECT smallfieldtogetareturn
FROM MyTable
WHERE MyTextField LIKE '%#mystringoflongdata#%'
</cfquery>
Then to see if it posted:
<cfif checkforexisting.RecordCount GT 0>
do the don't clip optional processing
</cfif>
What is killing me is that it finds the match SOMETIMES. Just
not ALL the time. I have tried no percent signs when passing the
ColdFusion variable, only one on the right (which has solved these
types of problems before).
If I pull the record back OUT of the database in a page by
itself (retrieving it by key field), it finds every other copy of
the code EVERY time.
I have put this into a stored procedure, which seemed to help
I even have resorted to creating a temp table first with the
clip, then running a query to retrieve the record that was just put
in, and comparing THAT to my primary table (both have text type
fields). The reason for this, is that if I write a single page that
has just the retrieval of the temp record, and that compare is run,
it finds EVERY copy of it in the primary table. I also thought that
maybe there was a strong problem (I had tried HTMLEditFormat,
HTMLCodeFormat, JavaCast since it was coming from a java program on
the browser end).
I am not terribly familiar with using large text fields, or
dealing with these large pieces of HTML code as something to
compare!
- Mike

Answer 1

MichaelSJudd wrote:
As Dan indicated your most likely issue is slight differences
in
capitalization and|or white space or other 'invisible'
difference
between your strings. One can have nearly infinite variety in
HTML code
that will display identically in a browser, the standard is
very
forgiving of formatting differences.
Something I have done in the past to compare HTML content is
to use the
hash() function. It will not help you eliminate the
differences but it
would show when something is different then it appears to be.
#hash(mystringoflongdata)# will return a hexadecimal number
representing
the string. Two *exactly* the same strings will produce the
same
number. Then it is very easy to compare these numbers to see
if two
long strings are the same.
To use this with your database data, you would probably have
to store
the hash value of the string in a field when you store the
HTML data.
HTH
Ian