Column names vs reference numbers

Language Help SQL
1

Why do I get different results between these 2 queries?

(A)
SELECT category, ROUND(AVG(valuation), 2)
FROM startups
GROUP BY 1
ORDER BY 2 DESC;

(B)
SELECT category, ROUND(AVG(valuation), 2)
FROM startups
GROUP BY category
ORDER BY valuation DESC;

The only difference is that the first uses reference numbers to the columns selected, the second uses the column names. Oddly, using the column names does not put both null values at the bottom of the list:

|Data Analytics | null |
|E-commerce |60250000.0|
|Video Streaming| null |

https://www.codecademy.com/courses/learn-sql/projects/fakeapps?action=resume_content_item

2

@cztechgmail.com
I had the same problem:
I first ran your (B) query, and its results are incorrect also in other places. For instance,

Education 2023800000.0
Enterprise 38508333333.33

so, 2,023,800,000 (2 billion) is larger than 38,508,333,333.33 (38.5 billion) , in this list.

However, the reason the result for (A) and (B) are different, is that in (A), 2 stands for AVG(valuation), not just valuation. Hence, if you use just valuation, the results are different.
A query like so:

-- (C)
SELECT category, ROUND(AVG(valuation), 2)
FROM startups
GROUP BY category
ORDER BY AVG(valuation) DESC;

gives the same results as your (A)