Use filters to specify the datapoints to include in the metric calculation for your validator. You can define filters for segmentations on a validator. The following table lists the types of filters that Validio supports:

For example, to count only the NULL values for a certain field, use a Volume Validator and apply a NULL filter.

📘

Note

A filter applied on a segmentation will be true for all validators that use the same segmentation.

Boolean Filter

Filter records based on a boolean field.

Enum Filter

Filter records based on whether field values are part of a defined set.

Null Filter

Filters records on whether a field values are NULL.

String Filter

Filter records on whether string field values match specific criteria.

🚧

Caution

Depending on source system, Validio leverages different regex syntax:

• For Data Warehouses and Query Engines, regex filters are applied in the source system. For more information on that syntax, refer to the regex syntax of your specific source system.
• For Object Storages and Data Streams, the regex syntax is compatible with the Rust regex library. For information on that syntax, refer to Rust - regex crate syntax.

Threshold Filter

Filter records on whether numeric values comply to a threshold.

*Specifies the 'allowed region'. For example,Greater than or equal than 5 specifies the allowed region, and all values strictly less than five are be filtered.

SQL Filter

📘

Note

SQL filter works for all Source types, including non-SQL data sources.

You define a SQL filter by typing a valid SQL WHERE statement in the text box, but omitting the word "WHERE". An example of a valid SQL filter is: some_column = 5.

Different data sources may use different SQL syntax. For more information, refer to the documentation of your data source. For non-SQL data sources, such as Data Streams or Object Storages, you find the supported functionality and syntax in the SQL filter expression reference.

SQL Filter Recipes

Below are several useful recipes for common SQL filters. Each recipe is written in a generic syntax, or in syntax specific for BigQuery or Snowflake, when required.

You can use these recipes as a starting point when writing SQL filters for other data sources, such as Redshift or Athena.

Filter operators

For example,some_column matches a filter with several operators.

some_column IS NULL OR (some_column > 3 AND some_column NOT IN(4,6))

Exists in set

For example,some_column match either of the values 'some_value' or 'some_other_value'.

some_column IN ( 'some_value', 'some_other_value' )

Missing values

For example,some_column is null or matches a value that represents a missing value.

some_column IS NULL OR LOWER(some_column) IN('', '0', '-', 'null')

Semi-structured data

For example,some_array contains at least one element with value 2.

2 IN(SELECT * FROM UNNEST(some_array))

array_contains(2, some_array)

String pattern

For example,some_column does not contain a valid URL.

NOT REGEXP_CONTAINS(some_column, '^(http(s)?:\\/\\/.)?[-a-zA-Z0-9@:%._\\+~#=]{2,256}\\.[a-z]{2,6}\\b([-a-zA-Z0-9@:%_\\+.~#?&//=]*)')

NOT REGEXP_LIKE(some_column, '^(http(s)?:\\/\\/.)?[-a-zA-Z0-9@:%._\\+~#=]{2,256}\\.[a-z]{2,6}\\b([-a-zA-Z0-9@:%_\\+.~#?&//=]*)')