StringPairReplacer v2 failed to upgrade to StringReplacer v6 

Thanks for the suggestions. I just read the documentation on StringPairReplacer. It says in Replacement Pairs, that if the strings contain a space, they must be escaped with a \. Instead of using a backslash, I enclosed my string with spaces in quotes. I also separated my pairs with new line characters (which isn’t mentioned in the documentation). My original Replacement Pairs has worked fine for years. For some reason, it only works correctly if the integer values are in descending order.

8 "Multiracial"
7 "White"
6 "Pacific Islander"
5 "Native American"
4 "Middle Eastern or South Asian"
3 "Hispanic/Latine(x)"
2 "Black/African American"
1 "Asian"

The StringPairReplacer upgrader (to StringReplacer) fails on the quotes and new lines. I discovered that if I change Replacement Pairs to the following (adding backslashes, and removing quotes and new lines) then the StringPairReplacer upgrader works fine.

8 Multiracial 7 White 6 Pacific\ Islander 5 Native\ American 4 Middle\ Eastern\ or\ South\ Asian 3 Hispanic/Latine(x) 2 Black/African\ American 1 Asian

I think that adjusting my Replacement Pairs for each StringPairReplacer before doing the upgrade to StringReplacer might be the best approach. This would minimize changes to my workspaces. And I wouldn’t have to create a bunch of Excel files.

Is there a reason for me to consider switching to AttributeValueMapper? If I were to write these workspaces again for the first time, I would likely use AttributeValueMapper. 

At first, I thought that AttributeValueMapper would be perfect, but I discovered that it maps the new value into a new attribute. StringReplacer replaces the new value in place (not a new attribute). My current logic is to replace the value in the same attribute. AttributeValueMapper could work, but I’d have to add and remove many attributes. Keep in mind, I have about 20 attributes similar to Ethnicity spanning about half dozen workspaces.

DatabaseJoiner seems like a good approach. I like the idea of storing everything in one Excel file, and presumably many sheets. But again, this would require adding and removing many attributes. And I don’t expect my values to change.

Side note: My source Ethnicity attribute actually is a comma separated list of integers (since a person can have multiple ethnicities), but I created a new attribute called EthnicityKey which is a single numeric value. If Ethnicity is a single value, I assign it to EthnicityKey, otherwise I assign the value ‘8’ to EthnicityKey meaning Multiracial. Then I assign EthnicityKey to Ethnicity to produce a single numeric value and remove the EthnicityKey attribute. If I was using AttributeValueMapper, I could avoid this and simply check for single numeric values and have Default Output Value of Multiracial.