How Do I Callout a Value

[Fescobar]

I created a Join on two columns and all of the data married-up accordingly.  I need to be able to call out two values that are accurately joined without removing them from the sheet.  In other words, I don't want to use the Unique tool to separate them.  I just need them call the difference out so that the viewer is able to see it easily.  

In the example below the call out I would like to do is on the Retailer Desc.  As you can see the join worked by matching the Corp UPC and UPC Clean columns.  I'd like to be able to do some sort of formatting, highlight the cells, change the font number...anything to help distinguish them.

Corp Item Nbr	Corp Description	Retailer Desc	Corp UPC	UPC Clean
142802	MOLLYDOOKER CAB SV GIGGLEPOT(SC)11	Mollydooker Cab Gigglepot Wine - 750 Ml	933697500020	933697500020
244379	WILD TURKEY AMER HONEY SHOT MACH	Mollydooker Cab Gigglepot Wine - 750 Ml	933697500020	933697500020
910914	MOLLYDOOKER VERDEL VIOLINIST(SC)17	Mollydooker Verdelho Violinist Wine - 750 Ml	933697500006	933697500006
958350	MOLLYDOOKER CAB SV GIGGLEPOT(SC)18	Mollydooker Cab Gigglepot Wine - 750 Ml	933697500020	933697500020
964419	MOLLYDOOKER SHIRAZ CARNIVAL LOVE18	Mollydooker Carnival Of Love Shiraz Wine - 750 Ml	933697500001	933697500001
987232	MOLLYDOOKER MERLOT SCOOTER(SC)19	Mollydooker The Scooter Merlot Wine - 750 Ml	933697500009	933697500009
987394	MOLLYDOOKER SHIRAZ BLUE EYED(SC)19	Mollydooker Blue Eyed Boy Shiraz Wine - 750 Ml	933697500030	933697500030

Thank you!!

FEscobar

[ArtApa]

Hi @Fescobar - You attached a workflow without data. Would you be able to provide sample data? 

[Fescobar]

Thank you .@ArtApa!

[CharlieS]

Hi @Fescobar 

I might approach this by highlighting the rows that contain the duplicate values in the [Retailer Desc] field. Here's how you can use some of the Reporting tools to change the fill color of duplicated values:

- Start by using a Summarize and Filter to determine which values are duplicated and need a color assigned.

- Then assign a randomized color to each value, so if you have 3 of the same value in your data, you can find all 3 more quickly. This process isn't perfect since colors could be duplicated or be very close, but it generally works alright. This could be circumvented by using a set list of color and I've included a container to show how that could be done as well.

- Build the color RGB values into the string format the Table tool can use like "rgb(194,241,253)"

- Join these values to the appropriate fields and configure a Row Rule in the Table tool to use that color value string like this:

Now your output file has the rows with non-unique values color coordinated like so:

An example workflow is attached, Let me know if this helps.

[Fescobar]

Thank you .@CharlieS

This is pretty cool stuff.  My requirement is only to call out  the differences in the Descriptions. Basically the Corp Description and the Retailer Description are different; however, the Corp UPC and UPC Clean are the same. therefore the union on both UPCs is accurate. 

From your screenshot, I would only need the values in Row 2 to be highlighted.  

Thank you SO much, again...this is cool stuff that you're sharing.

[CharlieS]

Ah I see that now. Thanks for clearing that up.

It would be easy enough to find records where the Retail Description has multiple Corp Descriptions assigned to it, but if we want to only highlight the "outliers" how do you imagine we could approach that? This sample is easy because there's two occurrences of "MOLLYDOCKER" and one of "WILD TURKEY", so maybe we just take the mode value as the correct match? What if the mode value is wrong, then some sort of fuzzy matching system might be necessary. 

I put together an example of using the model to determine the mode value and use that. The only problem is when there are even counts of each value, then the default is to use the first one. 

In this scenario, 'MOLLYDOOKER CAB SV GIGGLEPOT(SC)11' would be used. Hopefully this doesn't come up much, but it's something to think about. 

[Fescobar]

Thanks for the quick reply .@CharlieS 

Let me take a look and see how this looks.  It doesn't have to be perfect, as long we get the call out on when the Item/Material Desc are way off.  A fuzzy match would be pretty hard to maintain, we have hundreds of thousands items.

[Fescobar]

.@CharlieS

I see what you're saying now...

Your statement of "so maybe we just take the mode value as the correct match?", this would highlight everything that is accurate and leave the ones that are not as is?  If so, this will work too, I have to admit, that I am not sure how to you created this workflow.  It's pretty awesome!

Thank you!!

Fausto

[CharlieS]

I'm glad I could help. If there's any part you'd like me to explain more, I'm happy to dive in. 

[Fescobar]

Thank you .@CharlieS

How can I change the mode to highlight the entries that don't match?  Another big obstacle is that the names that my org uses will almost never be the same that the retailers use.  This is a big obstacle, so I just need to call out the ones that WAY-OFF.  Like the example above "MOLLYDOCKER" to "WILD TURKEY".  Any ideas on how to skin this?