Query
Columns visualizations are powered by its query engine. Most queries consist of:
- Metrics: Values or aggregated values from your dataset
- Dimensions: Keys used to break down metric values
- Filters: Conditions used to narrow down your data
Different people are familiar with different terminologies, we try to use the most common terms in the industry. Just to be clear: Metrics/Values are numbers that calculatable from your dataset, Dimensions/Keys are the columns that you can group by or break down your metrics.
Metric (Value)
A metric is a numeric field, typically used to measure or aggregate data.
You can configure metrics to compute:
- Count of distinct values in a column
- Row count
- Numeric aggregations, such as:
- Sum: Total value of a column
- Average (avg): Mean value
- Min/Max: Minimum or maximum value
- Percentiles: e.g., 25th, 50th, 95th percentile for a column
Dimension (Key)
A dimension is usually a text or categorical field used to group metrics. You can break down your metrics by any number of dimensions.
A powerful feature here is pivoting, which reorganizes your data by turning values from one dimension into columns while grouping by another. The pivot function is only available when:
- There is only one metric selected
- There are multiple dimensions
When pivoting by one dimension, all other dimensions are grouped together as combined dimensions.
Columns try to automatically detect the data and provide optimal query for you. Please pay attention to Keys section to check if special query is available for your data, eg. "Pivot All".
Filter
Columns supports two types of filters:
- Query filters: Applied before data is processed, affecting metric and dimension results.
- In-graph filters: Interactive UI controls embedded in the chart; users can toggle them on or off to adjust the view.
Don’t confuse the two: query filters shape your dataset before visualization, while in-graph filters help explore data dynamically after rendering.
Supported Query Filter Operations:
Most operations align with standard SQL. Wildcards like % can be used for pattern matching.
- = (equals): Field must equal a value
- != (not equals): Field must not equal a value
- IN: Field must match one of multiple values
- NOT IN: Field must not match any of the specified values
- IS EMPTY: Field is null or empty (for strings)
- IS NOT EMPTY: Opposite of above
- LIKE: Field matches a pattern (e.g., %value%)
- NOT LIKE: Field does not match a pattern
- ILIKE: Case-insensitive pattern match
- NOT ILIKE: Case-insensitive pattern mismatch