Index column for dimension table?

[ModelFear]

Hi all, 

I thought I post this question as it is so trivial that likely no one would ask it. 
I would like to create a simple dimension table.
I watched a couple of youtube videos and they added an index column for each dimension table. I identify with vendor names better than numbers and wonder if the index column will be confusing.

As a beginner, naturally I asked Ai to be sure, and I got this answer :

Q: In power bi, if i create a dimension table for vendors from a large table of purchases from vendors, why do i need to create an index column? Can't i just use the vendor name?
A: You're right, in Power BI, an index column in a dimension table (like vendors) typically doesn't improve query performance like it does in a database. Here's why vendor name might be sufficient and when an index column can be helpful:

Using Vendor Name:

Vendor name is often the natural key for the dimension table. It uniquely identifies a vendor and is easy to understand for analysis.
Power BI's in-memory engine is optimized for working with columnar data. It can efficiently search and filter based on the vendor name itself.

Index Column:

An index column can be useful in specific scenarios for data manipulation within Power Query Editor (PQE).
It can help with:
Numbering grouped data: Assigning a sequential number to each row within a group (e.g., ranking vendors by total purchase amount).
Identifying specific rows: Referencing rows by their index for custom logic transformations in PQE.
However, there are downsides to using an index column:

It adds an extra column to your table, increasing file size (though usually minimal for vendor tables).
It can potentially slow down query folding, a process where Power BI pushes calculations to the data source for better performance.

So should i create an index column? 😞

Thank you very much

[rajendraongole1 · Accepted solution]

Hi @ModelFear -Using an index column for a dimension table is generally a best practice if we check the good example that i came across:

1. Unique identifier for each row in a table when you referenced with fact table relationship

2.Good for performance optimization for identifying the distinct record in your dimension table.Indexes are typically faster to join on because they require less memory and processing power.

3. Index columns help to avoid issues related to null or duplicate values in natural keys (like names or dates). An index is always unique and not null.

Did I answer your question? Mark my post as a solution! This will help others on the forum!
Appreciate your Kudos!!

[ModelFear]

Hi rajendraongole1 

Thank you! I will follow this neat 1 min video to hide the index column after creating it. 
https://www.youtube.com/watch?v=dN3QYYbdk9c&t=1s