Tutorial Discussion: Migrate Target from at.js to Web SDK tutorial

[dwright]

Use this thread to ask any questions related to the Migrate Target from at.js 2.x to Platform Web SDK tutorial on Experience League. Experts are monitoring this thread to ensure your questions are answered.

[SachinJ]

Hi @dwright ,

Here is my response:

• Are you sure there is no issues with your feed? 
  • I don't think so as I have followed the methods of uploading feeds via .csv from here: (I am attaching sample feed in the response). And after successful feed I see all details filled in accurately as per the feed details "https://experienceleague.adobe.com/en/docs/target/using/recommendations/entities/feeds" 
• Are you able to see some records from your feed that are intact?
  • Yes, Initially when there is no corrupt feed update, all the product catalog looks good without loss of data, but as and when time passes and more mbox events flow in PDP, catalog starts getting updated with only entity.id retained rest is overwritten with blank value
• Are you sure you have escaped the fields in your field that need to be escaped?
  
  • I am not 100% sure here if I understood this correctly, but here is what I have done, the field that are blank in feed I have left it with "blank space" as suggested in the above link, coz without blank space feed does not work while uploading. I think am missing the escaped part ? Can you elaborate more on this please. Also, I am confused because instead of corrupting all the fields it overrides all except entity.id.. and I see entityId as is.(Screenshot attached)

[SachinJ]

Hi @dwright ,

Any suggestion if you review the above comments?

[dwright]

Hi,

So, there are two main ways to populate the catalog:

1. Feeds
2. Page calls

When using feeds, the page calls on product pages should populate the entity.id as that is a valuable key to use for recommendations. But calls containing the entity.id would only overwrite or supplement what was in your feed if it contained values for other entities. If the page call passed the entity.id with the correct value and also entity.name with an empty string, that would be one way to overwrite the value from the feed. That would be an easy thing to rule out.

But also, check this. In the screenshot of your catalog entry the entity.id is "07675062185" but none of the entity.ids in your csv contain a leading zero. What do you see if you look up the catalog entry without the leading zero?

Another thing to check is your environments. Your csv contains "stage" values. Could it be that you are populating the catalog correctly in your Recs staging environment catalog, but you are looking up the entities in a different environment which is separate from the feed?