Table Recipe: Data quality with semantic types

Table Recipe now has native data quality capabilities to detect, visualize, and fix invalid or inconsistent data. These capabilities make it easier to prepare clean, reliable datasets for analytics or machine learning, without leaving the no-code, spreadsheet-like interface.

• Data type and semantic type discovery: Each column is automatically assigned a data type based on the data it contains. Data types include native data types such as integer, date, and text, as well as semantic types: predefined (email, phone number, postal code, country code, and more) or custom (dictionary-based, pattern-based, or compound). You can manually change the assigned type at any time from the column menu.

• Column quality bar: Each column header displays a quality bar showing the proportion of valid (green), empty or null (black), and invalid (red) values.

• Cell-level invalid value indicators: Cells containing invalid values are flagged in red.

• Validity-based filtering: Click any segment of a column’s quality bar to instantly filter rows by quality status (valid, invalid, or empty or null). Filters can also be applied from the column menu or the filter dialog.

• Cleanse and contextual remediation suggestions: The Cleanse function lets you replace invalid, null, or empty values in a column. Table Recipe also surfaces relevant remediation suggestions, such as replace, clear or fill, based on detected invalid values, as well as conversion function suggestions based on the detected type.

Data quality features are available in both standalone table recipes and the table recipe processor in a data flow, and requires the data quality license entitlement (Qlik Cloud Analytics Premium and Enterprise, and Qlik Sense Enterprise SaaS).

Assessing the quality of the data