Filtering narrows down the data available in a workbook to aid focused analysis and to help answer specific business questions.

In Astrato the filtering capabilities help you create visual data stories and experiences with ease.

Astrato's Semantic Layer offers the option to apply a filter across all tables (applying a join), even if that table is not required in analytics/visualization result data. This extends the data reduction capability by always checking against data reduction lookups, to ensure users are always served the most relevant data.

You can enable this by using the "Apply across joins" option, on any table which has a filter applied.

Global filters apply across the entire workbook session for each user. If a user selects a year, all data in the workbook reflects this - unless a chart filter or measure filter has a rule that takes precedence.

Want to set an object to look at specific criteria, like a date range or region - then chart filters are what you are looking for! Chart filters enable setting permanent filters for specific objects. This is accessed from the Filters tab.

Each filter has a plus sign (+) beside it, click the + to add more filters to an object.

To add chart filters to an object, select the field, operator, and value. Once completed, click the Apply button to confirm. Repeat the process with additional measure filters.

Thinking of creating a comparison between measures containing different filter criteria? Measure filters are what you need. Compare sales between listed partners, and business units, all in measure filters, without needing to create new fields or groupings in your source data.

The example below shows revenue in two ways the first is revenue from won deals, and the second is potential revenue, from deals expected to be won this month.

The number in brackets next to the measure filter title indicates the number of measures that have filters applied.

With parameterized views, you can build dynamic, user-driven queries without rewriting SQL each time.

Custom SQL statements can be created with parameters inside your view. When the view runs, Astrato swaps in the actual values — from users, variables, controls, or defaults — to generate the final query.

How it works

Example

A view with source:

Parameters:

This allows one reusable view to power many scenarios — filters, what-if inputs, user-personalised data slices — all without duplicating SQL.

In each object, you can link a variable (or hardcode a value) to each parameter you have in use.

The filter bar is enabled automatically on new workbooks. To show or hide the filter bar, go to Workbook settings and use the field to select to hide or show the filter bar.

The filter bar can be styled to suit your workbook background. If you've set an image or color on your worksheet, the filter bar can blend in seamlessly by adjusting the Opacity or Background blur with an image or color on your worksheet.

The filter bar visible at the top of the worksheet is now enhanced with editable fields to simplify and speed up data analysis for you, and the teams you share your data with. The drop-down lists have a check mark beside the filters applied to the data object.

Select the filters you want, and a checkmark displays which filters are active on the data object.

The filter object enables end-users to quickly narrow down the global filter state across the workbook.

Measures and progress bars help visualize the general make-up of your data, helping you select right-first-time filters on the data that matters.

ℹ Filters are not synchronized across browser tabs

Tip: Enable Quickselect to bypass the apply/cancel buttons.

The date picker object enables users to select a range of dates easily.
​

The slider filter is great for continuous numeric fields; users can easily pick a range of values. Furthermore, by showing the distribution of values, users can focus on the meaningful parts of the data.

The image shows a slider with a selection made. The overall distribution is visible, and the selected range is highlighted in blue.

This feature lets you lock important filters so they always have exactly one value selected. It prevents reports from ever falling into an “unfiltered” or “multi-select” state.

Locked Filters give report designers more control over how people use dashboards. They let you fix important filters in place so viewers can explore data safely without changing key context.

You can mark specific fields in a workbook as Always One Selected.

Once enabled:

This keeps key business context locked in place.

Workbook Panel → Filtering section → Always One Selected Fields

Two inputs are used:

It uses the same style component as “Hidden Fields” for consistency.

This makes it obvious which filters are mandatory.

Use this for:

Filters in Astrato are not one thing. Each kind does a different job, and each one maps to a different SQL clause. Knowing which filter to reach for, and where it sits in the query, is how you build fast, trusted reports without waste.

There are four places a filter can live in Astrato's generated SQL:

Workbook filters, the filter bar, locked filters, date pickers and sliders all feed into one of these. They are how users set the values; the four clauses above are where those values land.

Runs at the lowest level. Shrinks the source table first. Then the join happens.

sql

With "Apply across joins" switched on, that same filter rides along every related table, so data reduction stays global.

Applied after the joins, on the rows the chart consumes.

sql

Two measures, two rules, one chart.

sql

For tests on the aggregate value, not on the rows.

sql

One business question, four filter types at work.

"For EMEA base data only, show me regions with total sales above one million on premium products sold in 2024, split by won and expected revenue."

sql

Read it bottom to top if you like: the semantic layer filter cut the rows first, the join brought customers in, chart filters shaped the slice, the measure filters split the aggregate into two views, and the result filter kept only the regions worth talking about.

Use this as a reach-for rule.

A short list worth pinning above your desk.

Answer: Semantic layer table filters. They sit at the base of the stack and run before any join. Because of this, they cut the data before Astrato stitches tables together, and every later step has less to carry.

Answer: Yes, and often a lot. A filter that runs early trims rows sooner. A filter that runs late still has to pull those rows into the join or the group before it can remove them. For large fact tables and fan-out joins, push filters down into the semantic layer wherever you can. The "Apply across joins" switch is built for this.

Answer: It lifts a single table filter so it rides along every query that touches a related table, even when that table is not in the chart. This extends data reduction to the whole workbook, not just one object, and keeps answers consistent across every view.

Answer: No. The two stack. Workbook filters set the stage for the whole session. Chart filters narrow one object further. A user pick at workbook level plus a rule on one chart combine; neither cancels the other out.

Answer: When two measures on the same chart need different rules. Chart filters narrow the whole object. Measure filters split the logic inside, so you can show "won revenue" and "expected revenue" side by side, each with its own gate, without adding new columns or groupings to your source.

Answer: A result filter is a rule on the aggregated value, not on the rows. "Show regions where total sales pass one million." That test only makes sense after SUM has run, so it belongs in HAVING. A WHERE clause would never see the total.

Answer: Chart filters apply after the join. If the join fans out rows, the filter still works, but your aggregates may not match what the business expects. The fix is usually a semantic layer table filter, applied before the join, so the rows that cause the fan-out never enter the mix.

Answer: Yes. Locked filters pin state. Always One Selected stops a field ever sitting empty or multi-select. Put together, they are ideal for customer portals, region gated reports, and single-tenant embeds where context must never drift.

Answer: No. Each tab holds its own filter state. Treat each tab as its own analysis.

Answer: Parameterized views push filter values into the SQL of a custom view at run time. The value is written into the query itself, not added later as a WHERE. Use them for OEM cases, multi-tenant apps, and what-if inputs where the SQL needs to change shape with the user.

Answer: No. The filter bar is a surface for setting values, not a fifth filter type. Values set there flow into the same workbook-level WHERE as any other global filter.

Answer: They combine as AND. A region chart filter and a workbook year filter produce one query with both clauses in the WHERE. If a rule on the chart contradicts a rule at the workbook level, the chart sees no rows, which is the honest answer.