SQL provides us with an easy way to automatically compute the moving average of a given column.
The moving average (also known as the rolling average, running average, moving mean (MM), or rolling mean) is a series of averages of different selections of the full data set. The moving average changes as we move through the data set. We can add the moving average to our SQL query result sets to see how it changes across the result set.
We can use SQL to create a running total of a given column. By this I mean, we can create a column that calculates the cumulative sum of a given column.
The running total/cumulative sum is the sum of all rows up until the current row. The result is that the sum increases (or decreases in the case of negative values) with each row returned in the result set. Each row’s value is added to the cumulative amount from all prior rows, so for any given row, we get the total sum for all rows up to that point – the “running total”.
In SQL, we can use the following method to compute the number of items within a given range when using a window function.
By this I mean we can use a frame clause in the window function to output how many items have a value that’s in a similar range to the value in the current row.
If you’re getting MySQL error 1140, which reads something like “In aggregated query without GROUP BY, expression #1 of SELECT list contains nonaggregated column ‘world.City.District’; this is incompatible with sql_mode=only_full_group_by“, it could be that you need to introduce a window function to the query.
This error can happen when we want multiple rows to contain aggregate values. For example, we might want to return subtotals of all rows within a group of rows. A window function can help us achieve this outcome.
When we use an OVER() clause to create a window function in SQL, we often use a PARTITION BY clause to partition the results. This can be handy if we want to do stuff like calculate subtotals.
But we can also use an empty OVER clause to calculate a grand total.
We might want to do this if we’re already using another OVER clause to calculate subtotals, but we also want a column to provide the grand total.
We can use the OVER() clause to create window functions in our SQL queries. A window function can be a useful tool that allows us to do things like compute moving averages, rank items, calculate cumulative sums, and much more.
Another common task for a window function might be to compute subtotals.
If you’re getting MySQL error 3582, which reads something like “Window ‘w1’ has a frame definition, so cannot be referenced by another window“, it’s probably because you’re referencing a named window that has a frame clause.
This can happen when we use the WINDOW clause to define a named window for a window function, and we include a frame clause in that window. When we then try to reference that window, MySQL tells us that we can’t, due to the frame clause.
Sometimes when we run SQL queries we might want to group the results by a given column. A common way to do this is with the GROUP BY clause.
But sometimes the GROUP BY clause isn’t enough.
Sometimes we want the results presented in a way that the GROUP BY clause doesn’t cater for. The GROUP BY is perfect if we want to collapse groups of rows into a single row, so that we can get an aggregate value for each set of collapsed rows. But sometimes we don’t want this. Sometimes we want to see all of the rows, as well as the aggregate values.
This is where the PARTITION BY clause comes in.
In MySQL, the WINDOW clause is an optional clause that we can use to create a named window. The named window can then be referred to from a window function.
Many SQL developers define their window functions directly in the OVER clause. But that’s not the only way to do it. We can also define them in a WINDOW clause, and then refer to them in the OVER clause.
When we define the window function in a WINDOW clause, we name it. When we do this, we can refer to that name from the OVER clause. This eliminates the need to include the definition directly inside the OVER clause.
When running SQL database queries, sometimes we need to use a window function in order to get the results we’re looking for. A window function is an aggregate-like function that enables partitioning and ordering of data within a result set.
The OVER clause is what enables us to create a window function.
The examples below demonstrate how we can incorporate the OVER clause in our SQL queries.