how to join tables in SQL

Estimated reading time: 4 minutes

When running an sql select query, most likely we may have a need to bring in data from other tables.

The reason for this , is that tables have a specific purpose for the data they store.

One of the most important features in a data architecture, is to avoid duplication.

This has the impact of only storing data on a table that is required.

As a result joining tables is necessary if you want to get all the information you need .

As an illustration on a customer purchase , the purchase details would be on the purchase table.

BUT the customer details would not be stored there. They would be stored on the customer table.

To put it another way a customer may make many purchases, but you only need their name and address etc. stored once.

The object therefore of joining tables is to use this structure , and avoid duplication over many tables .

For example in the above table, if you had the customer name and address on the purchases table and customer table,

they would always have to be in sync, and it would bring up a maintenance headache for the database administrator.

So what was developed was primary and foreign keys, which helps join tables, based on unique similar values in a column in each table.

What are the types of joins that can be used?

Inner Join: An inner join is where you join two tables but only return the rows that match in both tables.

Left Join: A left join is where you return all the records from the left table, and any matched records from the right table.

Right Join: A right join is where you return all the records from the right table, and any matched records from the left table.

Full Join: A full join uses the left join and right join, and returns all the rows from both tables in the output.

Equally important, where there is no value found a NULL value will be returned.

Lets take some examples to explain the above concepts.

So we have two tables below:

Sales Database, Customer Table

Sales Database, Sales Table

Inner Join

If we run this code, we will return the below. For this reason the purpose of an Inner Join is to only return matched records common to both.

select a.customer_no, a.customer_type, b.INVOICE_NO 
from dbo.CUSTOMER a -- left table
inner join sales.dbo.sales b -- right table

Left Join

With a left join of the below code will return four rows. All the records from the left table and any matched records from the right table.

select a.customer_no, a.customer_type, b.INVOICE_NO 
from dbo.CUSTOMER a --left table
left join sales.dbo.sales b --right table

Right Join

With a right join of the below code will return three rows. All the records from the right table and any matched records from the left table.

select a.customer_no, a.customer_type, b.INVOICE_NO 
from dbo.CUSTOMER a --left table
right join sales.dbo.sales b -- right table

Full Join

Finally in order to run a full join, the result of the below code would be as per the below. The objective is to return all rows matching and unmatching.

In the output, there is a Null Value, meaning that that row has no value for INVOICE_NO which is a primary key.

In essence primary keys are not allowed to have null values.

select a.customer_no, a.customer_type, b.INVOICE_NO from dbo.sales b
full join sales.dbo.customer a

How to use wildcards in SQL

Estimated reading time: 6 minutes

Following on from from our previous posts on SQL, this post will help to explain how to use wildcards in your query.

What would you use a wild card in the first place?

When a data analyst dealing with a large dataset, it is most likely that they will not know every piece of data.

As a result data will come from multiple sources and will be in different formats.

Using SQL wild cards will aid the programmer in been able to get specific pieces of data that may cause data quality errors.

Due to your query’s nature, you may not know where the problem is , the answer is to use wild cards for their ease and flexibility.

So lets look at a data set and start to apply some of the logic above , to a practical example.

We are going to use SLlite again, below is the table we are going to run our query off.

As you will see we have three columns with data in it, the examples below will work off the “name” column.

Name has a number of data points that are quite similar, so lets start showing you how to actually use the wild card.

Filter the data for all values before the last value using %l

The output below is basically going to the name column only and asking it to return values , that have “l” at the end.

What the SQL is instructed to do is to look at each string, and where there is an “l” at the end, and characters before it, then return those records.

This is what using wildcards does, the % basically is saying give me any value before “l”, which has to be at the end.

As none of the values have “l” at the end it returns blank, which is correct.

If we rerun this , with %y, we get four values returned:

Filter the data for all values that start with A%

As a follow on from above say you want to find records that begin with A, but you don’t know what comes after the “A”?

Below, correctly it returns only three, and it is not concerned what comes afterwards.

Filter the data for all values where a “g” is in the middle?

In the above we looked at the start and end points of the string, and it return records that matched the criteria.

There maybe a scenario where you want to look for records, with a particular value that may be in the middle of the string.

In this example, we know that “g” occurs at the fourth position, so it will return all records where g is in that position, regardless of what is on either side.

In applying %%% it is basically saying return anything, if the fourth character which is g, irrespective of what is in the previous three characters.

Filter the data for all values where there is a space in the record

There are going to be records that have spaces in them, and sometimes that may or may not be wanted.

In order to find those records, we would apply the below wildcard in the SQL

Filter the data for all values start with an “H” and end with a “y”

In a dataset, you may want to find records that begin and end with specific values, but you are not sure or bothered what is in between.

Below we have changed the “%” for “_” in the query. This change allows us to ask for a start and end character.

Something to note, between the “H” and “y” there are five underscores (_) in there. Each one represents the no of values between the first and last character. If the string was only three letters long, then you would use one _ and so on.

Summary and conlusion

In this post, we have described what a wild card is and its uses. They are very handy for searching for a combination of value or values when you are not sure what else is in the string.

This is quite commonly used in pattern searching, and in data cleansing , most systems would incorporate it especially if automating tasks , it allows clean data to process without it coming to a stand still.

On our YouTube channel you can subscribe to find out more information about data cleansing, SQL and lots of different tips and techniques, below is the video for this post:

A list of wild card operators are as follows:

Wild cardDescription
%Either before or after a character, represents any character that could appear but is unknown.
_This is a single character of any value that may appear in a wildcard search, represents a space between characters.
^Inside brackets beside characters, tells the program to not return those characters.
Inside a bracket and between characters, represents the range to be found of the characters it is in between.
[]If you place characters inside this bracket, it requests the program to return any of those characters in the output.

We have lots of posts on this website that will help you build your data analytics skills.

how to select columns with SQL

Estimated reading time: 2 minutes

Building on how to select all records with SQL , you may want to learn how to get specific columns of data from your table.

Using SQLite again , we will now look to select two columns from the table below, namely name and iso_code.

The purpose of completing this, is to allow the user to start to only choose the data they want.

As a database becomes bigger and more complex, so does the ability to retrieve the data.

Furthermore the speed and turnaround time becomes more critical.

What is more if these queries are been used for real time applications:

  • These results need to be precise to what is needed.
  • Additionally the speed of them returning back the records needs to be quick.

Subsequently below is the query run off the above table that we will use:

As you can see it has returned the two columns we asked for, excluding the description column.

To conclude, one thing to note this returns all records, that currently exist on the table for those two columns.

To see a video tutorial on this, watch the full video below:

how to select all records with SQL

Estimated reading time: 2 minutes

As part of working as a data analyst, in order to retrieve your data you will most likely need to connect into a database, and look at the database records and then start analysing, but how?

At the start of any data analytics project you will need to understand the structure of your data and what are the issues within it.

First of all you will need to connect to your database, and then open an SQL editor.

In this example we are going to use SQLite, it is very popular and can be downloaded from here .

Once you have this downloaded, you will need to create your database, tables and records.

Today we are going to retrieve records from a table that I have already created, as per the screenshot below

So that is the able above, but how would I use SQL to retrieve all its records?

In esseence it is very straight forward, all you need to do is run the following code:

select * from

and it will provide you with the following output

and there you go, now you have retrieved all the records from the database table!

To see a video tutorial on this, watch the full video below: