Hello everyone. In this article, we will learn how can we pivot rows to columns in the Hive.

Pivoting/transposing means we need to convert a row into columns. We need to do this to show a different view of data, to show aggregation performed on different granularity than which is present in the existing table.

Consider you have following data from some company. It shows how much money they spent on each resource in each quarter for some project in that company.

input data

From this table, you want to show data like this.

expected result

We will see how we can use CASE statements and COLLECT_SET to get these result from an existing table.

Step I:

First, we can use case statements to transpose required rows to columns.

This query will you give you result in the following way.

step 1 result

We can see that there are null values and each resource is having multiple records. This is expected as we have not grouped this on resource id yet and for each resource, there may not be an entry for each quarter.

Step II:

In this step, we will use COLLECT_SET and group records on basic of resource id to get the desired result.

step 2 result

This is almost like our desired result.

Step III:

We can prettify our results in step two by modifying our query with CASE statements. This step is optional and it depends on in what form do you want to achieve your final results.

This will be our output.

expected result

AS we can see, by using CASE and COLLECT_SET we can easily convert / pivot rows into columns in Hive.

You can learn more about COLLECT_SET in Hive at COLLECT_SET AND COLLECT_LIST IN HIVE

Leave a Reply

Close Panel