How to Group by Year in SQL

Problem:

You want to group your data by year in SQL.

Example 1:

One of the columns in your data is transaction_date. It contains a date. You would like to group all your data by year and calculate the total money earned each year.

The data table looks like this:

Solution 1: Displaying the year and the money earned

SELECT
  EXTRACT(year FROM transaction_date) AS year,
  SUM(money) AS money_earned
FROM data
GROUP BY EXTRACT(year FROM transaction_date);

The result is:

Solution 2: Displaying the complete date, the year, and the money earned in the corresponding year

SELECT
  transaction_date AS transaction_date,
  EXTRACT(year FROM transaction_date) AS year,
  SUM(money) OVER(PARTITION BY EXTRACT(year FROM transaction_date)) AS money_earned
FROM data;

The result is:

Discussion:

In this example it's assumed that you don't have the year column. Instead you have the column with complete dates.

First, you need to retrieve a year from the date. You can use the EXTRACT(part FROM date) function to do it. In your case, you'd like to extract the year, so the part is year. The date is the column which contains the dates – the transaction_date column. It's a good idea to rename the column to year afterwards.

If you'd like to display just the year and the total money earned in this year, you can use a GROUP BY. The first selected column is the year extracted from the date. The second column is the aggregate function SUM(money). At the end of the query you need a GROUP BY EXTRACT(year FROM transaction_date) or, simpler, GROUP BY 1 (since EXTRACT(year FROM transaction_date) is the first column.)

If you'd like to display more columns, you need a window function (Solution 2). After SUM(money) you write the OVER() clause and, since you'd like to calculate the sum for each year, use PARTITION BY EXTRACT(year FROM transaction_date) inside it. Note that you don't yet have the year column when calculating the sum, so PARTITION BY year won't work – you would get an error 'column "year" does not exist'. You can read more about the window functions in this article.

Example 2:

One of the columns in your data is year. You would like to group all your data by this column and calculate the total money earned each year.

The data table looks like this:

Solution 1: Displaying the year and the money earned

SELECT
  year,
  SUM(money) AS money_earned
FROM data
GROUP BY year;

The result is:

Solution 2: Displaying year, month, day and the money earned in the corresponding year

SELECT
  year,
  month,
  day,
  SUM(money) OVER(PARTITION BY year) AS money_earned
FROM data;

The result is:

Discussion:

In this example it's assumed that you already have the year column.

If you'd like to display the year and the total money earned in this year, a simple GROUP BY is enough. If you don't feel comfortable with the concept of GROUP BY, take a look here where we explain it. You simply use the aggregate function (here: SUM) with the correct column and at the end of the query you group by year. You can rename the column using the AS keyword with a new name.

It's more complicated if you'd also like to display some other columns. Then you need the solution using a window function (Solution 2). You should use the aggregate function with the appropriate column (here: SUM(money)) and write the OVER() clause afterwards. In this clause, you should use PARTITION BY with the column by which you'd like to group. That's how you get:

SUM(money) OVER(PARTITION BY year)

In this solution, you don't use a GROUP BY clause.

You can read more about the window functions.