Question 1

Pulling one column only

Accepted Answer

SELECT name fetches only the name column from every row. The result has one column and three rows - one per employee. Selecting a single column returns just that column's values, not the whole row.

Question 2

What star really promises

Accepted Answer

SELECT * returns every column in the order they appear in the table's CREATE TABLE definition. There is no sorting, no NULL filtering, and no extra columns. Avoid it in production queries - adding a new column later will silently change what your query returns.

Question 3

Who controls the column order?

Accepted Answer

The SELECT clause controls the order of columns in the result - not the table's physical column order. Because price is listed before name, the result row is 10.00 | Widget, not Widget | 10.00. This trips people up when they assume results mirror the table definition.

Question 4

Compute on the fly

Accepted Answer

You can write arithmetic expressions directly in SELECT. Here quantity * unit_price evaluates to 3 * 15.00 = 45.00 for the row where id = 1. The result column has no name unless you add an alias with AS. Expressions can use +, -, *, /, and parentheses like in any programming language.

Question 5

Give the expression a name

Accepted Answer

AS renames the result column. The value 90000 * 1.1 = 99000.0 is returned under the column header new_salary. The alias affects only the output column name, not the calculation itself. Aliases are most useful for giving computed expressions a readable name.

Question 6

Cut the result at two

Accepted Answer

LIMIT N restricts the result to at most N rows. Without ORDER BY, the database may return any N rows - in practice most engines return the first N in storage order. Here the result is Alice and Bob. Always pair LIMIT with ORDER BY when you need a predictable subset.

Question 7

A constant meets a table

Accepted Answer

Selecting a literal repeats that value for every row in the source table - one output row per source row. With 3 employees the result has 3 rows, all with the value 'active' in the status column. This pattern is useful for tagging all rows from a table before combining them with UNION.

Question 8

Does WHERE trim the columns?

Accepted Answer

SELECT * always returns all columns defined on the table, regardless of which column you filter on in WHERE. The WHERE clause decides which rows appear, not which columns. All three columns - id, username, email - show up in the result.

Question 9

Four ways to name a column

Accepted Answer

The AS keyword is optional when defining a column alias. Writing the alias name directly after the expression works in all major SQL databases. The = and : forms are not valid standard SQL (though = is specific to some T-SQL syntax). Using AS explicitly is recommended because it makes the alias easier to spot when reading the query.

Question 10

Zero placeholder column

Accepted Answer

Mixing a column reference with a literal produces the column value alongside the literal for each row. Every row gets the constant 0 in the discount column, just as it gets its own name value. This is handy when you want to add a placeholder column before a join or union fills it in with real data.

Question 11

Distinct pairs, not distinct values

Accepted Answer

DISTINCT removes rows where the entire selected combination is identical. Alice + shipped appears in rows 1 and 3 - that pair collapses to one row. Alice + pending and Bob + pending are each unique, so they remain. The result has 3 rows. A common mistake is assuming DISTINCT applies per-column independently - it applies to the full combination of selected columns.

Question 12

Skip one, take two

Accepted Answer

OFFSET skips the first N rows before applying LIMIT. OFFSET 1 skips Alice, then LIMIT 2 takes the next two rows - Bob and Carol. Offset-based pagination is easy to implement but slow on large tables because the database still scans all skipped rows. Keyset pagination (WHERE id > last_seen_id) is the scalable alternative.

Question 13

Alice once, not twice

Accepted Answer

DISTINCT eliminates duplicate values from the result. Alice appears in two rows but SELECT DISTINCT customer returns her only once. The result has one row per unique customer value. Note that DISTINCT applies to the entire selected row - adding a second column means a row is only a duplicate if all selected column values match.

Question 14

Which clause can see the alias?

Accepted Answer

Column aliases defined in SELECT are only available to ORDER BY in standard SQL. WHERE, GROUP BY, and HAVING are evaluated before SELECT in the logical query order, so they cannot see aliases defined there. Some databases extend this - MySQL allows aliases in GROUP BY and HAVING - but relying on that behavior is not portable.

Question 15

The alias that WHERE never met

Accepted Answer

The logical order of SQL clause evaluation is FROM -> WHERE -> GROUP BY -> HAVING -> SELECT -> ORDER BY. Because WHERE runs before SELECT, the alias new_salary has not been defined yet when the database evaluates the filter. To filter on the computed value, repeat the expression: WHERE salary * 1.1 > 100000.

Question 16

No table, no problem

Accepted Answer

PostgreSQL and most modern databases allow SELECT without FROM when you are selecting an expression or function call - no table is needed. SELECT 2 + 2 returns a single row with the value 4. SELECT NOW() and SELECT gen_random_uuid() follow the same pattern. SELECT name fails because name has no table context. The other two options are syntax errors.

SELECT Basics