Comments on: The Fuzzy Lookup Transformation

By: BI Monkey

BI Monkey — Mon, 12 Dec 2011 02:21:05 +0000

@Mai – use an OLE DB transformation to update the target row

By: Mai Elsayed

Mai Elsayed — Sun, 11 Dec 2011 14:05:34 +0000

Please can anyone tell me how i update records in the refrence table depend on their matches on the source table ?

By: BI Monkey

BI Monkey — Tue, 04 Oct 2011 21:36:36 +0000

ashu.. if the company name data is reliable use a lookup. If it’s inconsistent use a Fuzzy Lookup.

By: ashu

ashu — Tue, 04 Oct 2011 06:05:54 +0000

hi..
i need to do same thing
plz help i have created one ssis package.. i want to insert price into table checking based on company name column …..

By: CozyRoc

CozyRoc — Fri, 29 Jul 2011 21:58:44 +0000

Just wanted to mention apparently there is a bug in the Fuzzy Lookup component, which appears on 64bit systems. I have lost a couple of hours banging my head with this. Please vote it is important for MS to fix this issue here:

https://connect.microsoft.com/SQLServer/feedback/details/488387/fuzzy-lookup-triggers-sqldumper-even-with-very-small-dataset-with-run64bitruntime-set-to-false

The problem is not new because I was able to reproduce it even under SQL 2005 SP3.

By: Darren Green

Darren Green — Mon, 13 Dec 2010 11:05:04 +0000

James, Setting WarmCaches to false is certainly worth trying when you have a large index and only a few rows to try and match. I’ve gained significant benefit on occasions by turning it off. Consider the ratio between the fuzzy index rows (the reference data) and the rows to match (the input rows), the larger this ratio the more likely it is that you’ll get a performance gain.

It is a bit like the traditional Lookup and the default cache option, it can sometimes spend more time populating the cache than it takes to just run the query direct.

By: Fuzzy Thinking | BI Monkey

Fuzzy Thinking | BI Monkey — Wed, 11 Nov 2009 00:19:34 +0000

[...] Fuzzy Lookup and Fuzzy Grouping transformations – guides to the practical use of the transformations from me, the BI Monkey [...]

By: BI Monkey

BI Monkey — Thu, 08 Oct 2009 22:52:57 +0000

Hi Kevin. I haven’t tested the Exhaustive property thoroughly but I found one case where a data set of about 100,000 rows jumped from about a 5 minute to 8 hour processing time when I switched it to True. My gut feeling is the threshold is pretty low, but because of the number of variables in any such process around hardware and match complexity I would just test in your environment and see how much of a drag it adds.

However what I did find is that for some matching scenarios, the additional matches it brought in weren’t all that valuable. How that would apply to your situation will depend on how well the matching algorithm will work with your data. Your best bet is to do a test on a sample of your data and see if the performance overhead is worth it in terms of additional matches obtained.

By: Kevin Moormann

Kevin Moormann — Thu, 08 Oct 2009 21:12:52 +0000

Have you determined a rule of thumb for how big a data set can be before performance becomes to big of an issue when exhaustive is set to true?

By: Pardeep Dhull

Pardeep Dhull — Sun, 30 Aug 2009 15:42:59 +0000

Hey thanks man….i ws looking for something like this…simple yet thorough one.