Group By Vs Distinct in SQL - sql

Question

SELECT continent, COUNT(name)
FROM world
WHERE population>200000000
GROUP BY continent
When i execute the query above the query runs fine. It basically shows the number of countries in each continent that has a population larger than 200000000.
However when I modify my query to the below :
SELECT DISTINCT(continent), COUNT(name)
FROM world
WHERE population>200000000
This does not work. I am wondering what the reason is. In this case I am saying for each distinct continent count the total countries with population larger than 200000000.
I just want to understand the reasoning so i can become better at writing queries.

Answer 1

Why does this not work?
SELECT DISTINCT(continent), COUNT(name)
FROM world
WHERE population > 200000000;
That is simple. You have an aggregation query, because you have COUNT() in the SELECT. You have no GROUP BY, so any other columns references in the SELECT must be the arguments of aggregations columns. So, continent generates an error.
You seem to also be under the impression that the parentheses around continent have some significance. They do not. Not at all. SQL has a construct, SELECT DISTINCT, which selects distinct values of rows.
Also note that DISTINCT is almost never used with aggregation functions.

Answer 2

I think you want:
SELECT continent
, COUNT(DISTINCT name) AS DistinctCountries
FROM world
WHERE population>200000000
GROUP BY continent
If want each row to represent a continent, you need to group by continent. Then count the distinct countries in the continent where your condition is met.

Answer 3

The first query and its order of evaluation:
FROM world: Get rows from the world table.
WHERE population>200000000: Only accept rows (countries?) with a population greater than 200000000.
GROUP BY continent: Aggregate the rows so as to get one result row per continent.
SELECT COUNT(name): For the continent show the count of its rows found in 3 where name is not null.
SELECT continent: show the continent.
The second query and its order of evaluation:
FROM world: Get rows from the world table.
WHERE population>200000000: Only accept rows (countries?) with a population greater than 200000000.
GROUP BY continent: Aggregate the rows so as to get one result row per continent.
SELECT COUNT(name): As there is no group by clause, this is saying you want one result row only, with the count of all rows found in 3 where name is not null.
SELECT (continent): The parentheses are superfluous. You are saying you want to show the continent. However, as you said with COUNT(name), you wanted to show one result row only, which continent are you talking about? It makes no sense to the DBMS and is invalid SQL. (There is one DBMS making an exception here, though: MySQL would just pick a continent arbitrarily rather than raising an error, a certain setting provided.)
SELECT DISTINCT: Of all result rows, you want duplicates removed, i.e. all rows showing the same continent and count.
Your error, as you can see, is in steps 4 and 5, where SELECT COUNT(name) without GROUP BY and SELECT (continent) don't match semantically.

Answer 4

GROUP BY AND DISTINCT are very much seperate in one way or the other.
Group by is used specifically to create and perform aggregation per groups while distinct is just used to have distinct/unique records or removing duplicates nothing else.
SELECT continent, COUNT(name)
FROM world
WHERE population>200000000
GROUP BY continent
The first query has a group by on continent it will group all rows which are having same continent into seperate groups after filtering via where.
This query will give you records of count per each continent
SELECT DISTINCT continent,
COUNT(name)
FROM world
WHERE population>200000000
The 2nd query means performing distinct and count on whole table but not groups (note) after filtering population. This query will give you distinct/unique continent but count is independent of groups and is of whole table