Table as a KB
# 🤝help
p
I am having problems with KB understanding my table. It has some trending stories' titles, descriptions, conversation sizes, dates, cities, and supporting URLs. But I have been trying to make use of it for the past week and changed the formatting of the table multiple times with no luck. What should I do? (Report ID: 71520428-0e35-42d5-b12b-a34a08b3847a)
f
Hello @purple-dream-91712 can you walk me through your use case please?
p
Hello Sara! I am trying to use to in order to archive/update trending stories and news alerts over a period of time. Then retrieve the most trending via the chatbot using queries like what's the most trending story in 2023..
f
How are quantifying trending?
p
There is a number of "trending" (200K) ..
f
Can you send me a screenshot of the table please?
p
Here is a part of it. I tried deleting _2 but it did not work
f
So Trending is being pulled from an API?
p
No its a number
f
How is the number the being created?
p
its a static number that is imported from a table
My problem is that I tried importing the same dataset into another table in another bot and it just showed null.
f
Okay so I am looking back at your screenshot from 6 March The trending column is not set to be searchable. How I know is the 123 next to Trending is not blue. Try making the column searchable, click on the 123 next to the name, toggle searchable to true, and click confirm, and let me know if you are still having the same problem.
p
Thanks a lot! Will try it
It worked but still, there is a problem with it sorting the stories based on trend or conversation size. when asked about most trending story in 2023 for example, it just pulls a random number. Is there any way to get it to understand the ranks and the trends?
f
Have you added descriptions to the columns in your table? Click on the symbol next to the name of the column and you can add in a description.
The other thing you can also do is work with Knowledge base query card and specify what to look at when here is a brief example of what I did that was listing all the topics in a table for a demo
p
Thanks a lot for these tips! They are super helpful, but there is a weird thing because it says "an interaction rate of 0.141", instead of the number in the coloumn
I think that's what is making sorting inaccurate
because the main problem is the random choice of stories although the numbers show how big is it
f
p
I am trying to implement that, but I am not sure I fully understand how will this serve the purpose. Because I need it to perform a relative comparison between stories at a given date.
f
The selector will be filter on your table so you can set the selector for a given date. It will in turn only return the pertinent rows and return that information. You can then use an AI task to run on the data returned from the selector.