Custom SQL Filter
Last updated
Last updated
CrossEngage's Predictions Platform provides various filtering options to add or remove Customers from certain Selections. This includes customer lists uploaded as files (blocklists and allowlists) or other Selections. These methods can be quite useful to restrict Selections, but can be time-intensive and limited in capability.
Hence the Platform allows you to enter a Custom SQL filter when creating Insights, Models or Prediction. For example, when creating a model, a Custom SQL Filter can be added in (8) to filter Target Variable, or in (10) to filter Target Group.
To use the SQL filter, write a SQL condition in the given field. The SQL Condition is the part of the query that can be written in a "WHERE" clause.
Target Variable: All fields contained in the transactional data
Target Group: All features automatically created or custom-built by yourself or a data scientist.
The fields can have the following data types:
Numeric: Whole numbers or floating point numbers
Strings: Any character such as digits, letters, special characters, etc. Strings are written in 'single quotes'.
Date & Time
Boolean: True or False
Operators can be used to compare or combine different values or expressions:
Arithmetic Operators: These can be applied to numeric values: +, -, *, /, etc.
Comparison Operators: These compare values ​​between two expressions and return a boolean value: =, >, <, !=, IN, BETWEEN, LIKE etc.
Logical Operators: Boolean operators, which can combine different expressions and values ​​together: AND, OR, etc.
Remember to use the correct case when using filters. A simple way to avoid case-sensitivity errors is to keep all values stored in the data tables in lowercase.
Wildcards are placeholder characters, which can be added to a string when a certain part of string is unknown or not fixed. Wildcards are used with the SQL LIKE operator to find similar strings.
Underscore: Underscore (_) can be used to replace a single unknown character in a string.
Percent: Percent (%) can be used to replace a series of unknown characters, of any length.
To Filter blue cups and green plates:
To Filter LD1, LD2 and LD3 Printers:
Alternately, we can use a Wildcard:
To Filter customers in France:
To Filter customers with recency between 10 and 20 weeks:
To learn more about SQL, Click Here.
