Join 2017_Deep Dive_Table Calculations 101
1 like623 views
1. Table calculations allow you to perform calculations on the fly within Looker reports to add new columns or metrics without modifying the underlying data model. They use Looker expressions rather than LookML. 2. The document provides examples of using mathematical functions like sum and rand, string functions like contains, and date functions like diff_days in table calculations to calculate new metrics like percentages of users by gender, average items delivered per week, and time to delivery. 3. Tips are provided like avoiding sorting on reports with many rows, limiting result sets before aggregation, and hiding table calculations from visualizations.
Join 2017_Deep Dive_Table Calculations 101
- 1. 1 Table Calculations 101 Nicole Beyer Manager Department of Customer Love
- 2. • What is a Table Calc? • Best Practices • How to create a Table Calc • Mathematical functions • String Functions • Date Functions 2 Agenda
- 3. • A way to create columns on the fly • Based on Looker Expressions (lexp) rather than LookML • Anyone can make them • Always based on the fields in an existing report 3 Overview
- 4. • You want to perform a one off calculation • You want to do something that would be really hard in SQL, like window functions 4 Best Practices
- 5. • Let’s calculate our total number of new male and female users per week starting here • We will use only basic + operator and no formatting • Now say we want to know what percentage of our new users are female • We will reference our previous table calc and use formatting • Final product 5 My first table calc
- 6. Mathematical Functions • Can be per row (round, rand) or aggregations (sum, max) which mean one result for all rows • Let’s use rand which operates per row to create a report which surfaces 5 random order items to review • Start here • Common pitfall: cannot sort on reports that have hit the row limit • Solution: Increase the limit or filter the results • Tip: Hide from visualization • Final product
- 7. Mathematical Functions Continued • Let’s use table calcs to take the aggregate of an aggregate • Say we want to know the average number of items we have delivered in the past 4 weeks • Start here with the count per week • Common pitfall: aggregations will only aggregate what is being shown on the report • Solution: limit the result set or do the calculation in LookML • Final product
- 8. • Allows you to manipulate strings • Similar to functions allowed in SQL • No regex at the moment • My favorite is contains • Let’s check this explore to see which product names contain the word ‘jean’ • Final product 8 String Functions
- 9. • You can add and subtract dates! • Similar to functions allowed in SQL but dialect agnostic (so nice!) • Let’s use diff_days to find the time that it takes for an order item to be delivered starting from this explore • Final product 9 Date Functions
- 10. • Take advantage of autocomplete! • Cannot filter • Can sort if not hitting the row limit • Will be grouped by any dimensions used in the report • Check out the docs! 10 Tips and Tricks