I personally follow the practice of testing my prompt directly with the LLM before setting it up in the enrichment column. When I gave your prompt to ChatGPT, it returned a Python-based explanation instead of the expected output. So, I added an extra line of instruction to your original prompt, and it worked well.
Here is the modified prompt:
I'm trying but the it return 2-leltter country anyway, here is the prompt I'm using:
#CONTEXT#
You are tasked with converting 2-letter country codes to their corresponding full country names using the ISO 3166-1 alpha-2 standard.
#OBJECTIVE#
Convert the value in the 'Country Enriched' column from a 2-letter country code to the full country name. If the input is invalid or empty, return "Unknown Country".
#INSTRUCTIONS#
1. Read the value from the 'Country Enriched' column.
2. Check if the value is a valid ISO 3166-1 alpha-2 country code.
3. If valid, return the corresponding full country name (e.g., "US" → "United States").
4. If the value is invalid or empty, return "Unknown Country".
#EXAMPLES#
Input: "US" → Output: "United States"
Input: "FR" → Output: "France"
Input: "DE" → Output: "Germany"
Input: "ZZ" → Output: "Unknown Country"
Input: "" → Output: "Unknown Country"
Don't give me any additional explanations or try to find mistakes with my input and use the above instructions to give me the answer to the below input in the desired format:
Input to parse: "AF"
Attached to this message is an image of the output received by ChatGPT.
.png){kind=logo}
