Converting data types

Reporting in SQL

Tyler Pernes

Learning & Development Consultant

Chapter goal

"Messy" data types

Issue 1: Type-specific functions

Issue 2: Combining tables (JOIN)

Issue 2: Combining tables (UNION)

Interpreting errors

Type-Specific Function Error:

SELECT AVG(first_name)
FROM athletes;

ERROR: Function avg(character varying) does not exist

JOIN Error:

SELECT country, continent
FROM countries AS c1
JOIN continents AS c2
ON c1.continent_id = c2.id;

ERROR: Operator does not exist: integer = character varying

Solution: Wrap it in a CAST()

Syntax:

CAST(field AS type)

Examples:

CAST(birthday AS date)
CAST(country_id AS int)

CASTing for functions

SELECT DATE_PART('month',birthdate)
FROM birthdates;

ERROR: Can't run DATE_PART on string.

SELECT DATE_PART('month', 
       CAST(birthdate AS date)) 
FROM birthdates;

+-----+
| 04  |
| 05  |
+-----+

CASTing for JOINs

SELECT a.id, b.id
FROM table_a AS a
JOIN table_b AS b
ON a.id = b.id;

ERROR: Cannot join ON varchar = int.

SELECT a.id, b.id
FROM table_a AS a
JOIN table_b AS b
ON a.id = CAST(b.id AS varchar);

Query Ran Successfully!

Planning for data type issues

Fix as they come up
Read error messages!

Planning for data type issues

SELECT column_name, data_type
FROM information_schema.columns
WHERE table_name = 'countries';

+--------------+--------------------+
| column_name  | data_type          |
|--------------|--------------------|
| id           | integer            |
| country      | character varying  |
| region       | character varying  |
+--------------+--------------------+

Data type documentation

https://www.postgresql.org/docs/9.5/datatype.html

Practice time!

Reporting in SQL