The Ultimate Oracle SQL Course

Filtering and Sorting Results

Section Recap

In this section, you learned that the WHERE clause allows you to filter the results of your que-
ries, and it is composed of the WHERE keyword and a condition that needs to evaluate to true
for the rows that are going to be returned by the query.

This condition can actually be composed of several sub-conditions, but at the end of the day, all
the sub-conditions are reduced to one single condition, which determines if the row has to be
returned.

The order in which the sub-conditions are evaluated is determined by Oracle based on the order
in which they appear and the operators' precedence, which is a topic that will be covered later in
the course, but if you want to force a specific order of evaluation of the conditions, you can use
parentheses to do it. The conditions that are nested at the deepest levels of parentheses are evalu-
ated first.

The order of evaluation is very important because as you saw in one example, a simple pair of
parentheses can make the query produce different results.

You also learned that even though you can write dates as if they were strings, it is better to use
date literals, because they work correctly on any system, regardless of the local date format set-
tings. Date literals are written using the word DATE, and a string that represents the desired date
in the format 'YYYY-MM-DD'.

Here is an example of how you can filter rows comparing dates correctly, using date literals:

SELECT *
FROM products
WHERE expiration_date < DATE '2016-06-01';

You can also create dates from strings by using a conversion function, but that is something that
will be covered in a future lesson.

Another method you have to filter the results of queries is by using the DISTINCT keyword,
which filters or removes duplicate rows, but remember: for a row to be removed, the whole row
must be equal to some other row.

© This is copyrighted material. Get the official course at sql.standout-dev.com 1

 The Ultimate Oracle SQL Course

Another very important topic that was covered in this section is the ORDER BY clause, which
allows you to order the results of your queries.

By default Oracle orders in ascending order, but if you want to order in descending order, you
just have to add the DESC keyword after the name of the column you are ordering by.

Also, you learned that you can order by more than one column, for which you just have to spec-
ify the desired columns in the order by clause, separated by commas, and you can use a different
type of ordering for each column.

It is VERY important that you remember that if you need or want the results of your query in a
certain order, you MUST add an order by clause, even if you appear to get the results already in
that order without adding this clause. If you fail to do that, the order is not guaranteed, and your
query could start producing results in a different order at any moment.

Instead of the name of the columns, you can also use column positions in the order by clause, so
if you order by 1, for example, it means that you want to order by the first column returned by
the query. This is very handy, but can sometimes make SQL code less readable.

The ORDER BY clause is very flexible, and you saw that you can actually use very complex ex-
pressions there, to do things like conditional ordering, and any other thing your creativity wants
you to do.

Here is a small example:

SELECT id, name, department_id, salary

FROM employee
ORDER BY
CASE department_id
WHEN 1
THEN salary
ELSE id
END;

In this case the ORDER BY clause includes a CASE expression, which is something you have
not learned yet, so, by now just remember that you use complex expressions in the ORDER BY
clause, without trying to completely understand the above example.

And finally, you learned about NULLs, so now you know that null is the equivalent to unknown,
and thus, you cannot compare it to actual values using normal comparison operators like equal

© This is copyrighted material. Get the official course at sql.standout-dev.com 2

 The Ultimate Oracle SQL Course

to, or greater than, because those comparisons result in an unknown logical value. The correct
way to compare NULLs or to write conditions to check for NULLs is the use of the IS NULL or
IS NOT NULL operators.

Keep in mind that a condition can evaluate to 3 different logical values: TRUE, FALSE, and UN-
KNOWN, and the only rows that will be returned are those for which the condition evaluates to
TRUE.

You should also think about nulls when ordering the results of your queries. By default, when
Oracle orders in ascending order, it leaves nulls at the end, and when it orders in descending
order it puts nulls at the beginning, before the non-null values, but if that is not what you want,
you can add the optional NULLS FIRST or NULLS LAST clauses in the ORDER BY clause of
your query.

There are some functions that simplify our work with NULLs, which includes:

• NVL, which lets you define a substitute for null values.

• COALESCE, which returns the first non-null value from the list of values or expressions
passed as parameters.
• NVL2, which allows you to return different values, depending on whether the first expres-
sion passed is null or not.
• NULLIF, which returns NULL if the two expressions passed as parameters are equal.
• And LNNVL, which is a little different because it receives a condition as a parameter and
is also used as a condition. You can think of it as an “IS NOT TRUE” condition.

NVL, NVL2, and LNNVL are Oracle-specific, and COALESCE and NULLIF are part of the SQL
standard.

Here is an example where these functions are used:

SELECT id,
name,
phone,
NVL(phone, '* No Phone *'),
COALESCE (salary, bonus, 0),
NVL2(phone, 'Has Phone', 'Does Not Have Phone'),
NULLIF(bonus, 100)
FROM employee
WHERE LNNVL(phone = '1.123.456.7890');

© This is copyrighted material. Get the official course at sql.standout-dev.com 3

 The Ultimate Oracle SQL Course

The above condition can be understood as “WHERE it is not true that the phone is equal to
'1.123.456.7890'”, so, it returns rows in which the phone is different from that number, but also
rows in which the phone is null.
See you in the next section!

© This is copyrighted material. Get the official course at sql.standout-dev.com 4