How Do You Filter the Top 10 in Power BI?

How Do You Filter the Top 10 in Power BI?

Almost every report eventually gets the same request: stop showing me all four hundred products and show me the ten that matter. It sounds like a sorting problem, so most people sort the visual descending, see the big values at the top, and consider it done. Then someone exports the table and gets all four hundred rows back.

Sorting changes the order of what is displayed. It does not change what is included. To show only the top ten you have to filter, and Power BI has a filter type built specifically for this that a surprising number of report authors never open.

This guide covers the built-in Top N filter first, because it solves the common case in about fifteen seconds, and then the DAX approaches you need when Top N runs out of road, which it does in three fairly predictable ways.

What Is the Fastest Way to Filter the Top 10 in Power BI?

Select the visual, open the Filters pane, drag the field you want to rank into the "Filters on this visual" area, change Filter type to "Top N", set Show items to Top and 10, drag the measure you want to rank by into the "By value" well, and select Apply filter.

That is the whole operation. The visual now contains ten items instead of four hundred, and every total, average and percentage in that visual is calculated over those ten only.

The part people miss is the "By value" well. The Top N filter does not assume you want to rank by whatever measure is already in the visual. You have to tell it explicitly which number defines "top", and if you leave that well empty the filter does nothing at all.

Not sure which reports are pulling their weight? Get a free SEO audit and see exactly where your traffic and conversions are coming from.

What the Top N Filter Actually Does

The Top N filter sits in the Filter type dropdown on a visual-level filter card, alongside Basic filtering and Advanced filtering. When you apply it, Power BI evaluates your chosen measure for every member of the field you dropped in, ranks those members, keeps the N highest, and discards the rest before the visual renders.

That last clause is the important one. The rows are gone, not hidden. Anything the visual computes afterwards sees only the survivors.

The Two Inputs: How Many, and By What

The filter card asks for two things. "Show items" takes a direction and a count: Top or Bottom, followed by a whole number. You can type 10, 5, 25 or 3 depending on what the question needs.

"By value" takes the field or measure that does the ranking. This can be the same measure the visual displays, which is the usual case, or a completely different one. Ranking products by profit while displaying their revenue is perfectly legal and sometimes exactly what the business wants to see.

Which Filter Level You Can Use It On

Power BI offers Top N only on visual-level filters, in the "Filters on this visual" section. You will not find it when you drop a field into "Filters on this page" or "Filters on all pages", because those levels have no single visual whose ranking context would define what "top" means.

In practice this is less restrictive than it sounds. If you want a whole page restricted to the top ten customers, you apply the Top N filter to each visual, or you build the restriction into a measure or a calculated table so that every visual inherits it.

Top N Is a Filter, Not a Sort

A sorted visual and a Top N filtered visual look similar at a glance and behave completely differently underneath. The sorted one still contains every row, so its totals, its exports and its tooltips all reflect the full dataset. The filtered one contains ten rows and nothing else.

Neither is wrong. They answer different questions, and the mistake is using the first while believing you have done the second.

Want to know how your site really performs in search? Run a free SEO audit and get a prioritised list of what to fix.

How Do You Apply a Top 10 Filter, Step by Step?

Here is the full sequence in Power BI Desktop, using a table of products and a revenue measure as the example. The same steps work for bar charts, column charts, matrices and most other visuals.

Step 1 - Select the Visual You Want to Restrict

Click once on the visual itself, not on the page background. The Filters pane is context-sensitive, and the "Filters on this visual" section only appears when a visual is actually selected. If you skip this, you will end up building a page-level filter and wondering why Top N is missing from the dropdown.

Step 2 - Open the Filters Pane

If the pane is collapsed, select the chevrons in the upper-right of the canvas to expand it. You should see up to three sections stacked vertically: filters on this visual, on this page, and on all pages.

Step 3 - Drag the Ranking Field Into the Visual Filters Area

From the Data pane, drag the categorical field whose members you want to rank, for example Product Name or Customer, into the area labelled "Add data fields here" under "Filters on this visual".

Use the dimension, not the measure. You are telling Power BI which list to shorten. A common misstep is dragging the revenue measure here instead, which gives you a numeric filter card with no Top N option at all.

Step 4 - Change the Filter Type to Top N

Expand the new filter card and open the Filter type dropdown. Select "Top N". The card changes shape and exposes the Show items row and the By value well.

Step 5 - Set Show Items to Top and Enter 10

Leave the direction as Top, or switch it to Bottom if you are hunting for the worst performers, and type 10 into the number box. The value has to be a positive whole number; Power BI will not accept a measure here, which matters later when you want the N itself to be dynamic.

Step 6 - Drag Your Measure Into the By Value Well

Drag the measure that defines the ranking, for example Total Revenue, into the "By value" well on the same card. This is the step that is easiest to forget and the one that makes the filter do anything.

Step 7 - Select Apply Filter

Unlike most filter cards, Top N does not apply as you type. Select Apply filter at the bottom of the card to commit it. The visual redraws with ten items.

If nothing changes, check three things in order: that By value is populated, that you dropped a dimension rather than a measure, and that you actually selected Apply filter.

When Should You Use the Top N Filter?

Use it whenever the question is genuinely about the leaders rather than the list. "Which ten products drive the most revenue", "which five campaigns wasted the most budget", "who are our twenty biggest accounts" are all Top N questions, and the built-in filter answers them without a line of DAX.

It is also the right tool for readability. A bar chart with four hundred categories is a grey smear; the same chart with ten is a chart. Restricting a visual to its top ten is often the single fastest way to make a cluttered page legible.

Top N behaves well inside a filtered page, which is its best quality. Because the ranking is evaluated in the visual's current filter context, a slicer set to a single region recalculates the top ten within that region. You do not get the global top ten filtered down to whoever happens to appear in the region; you get that region's own top ten. That is almost always what the reader assumes is happening, and here the assumption is correct.

The place to stop and think is when the number ten itself is doing analytical work. If the business question is really "which products make up 80 percent of revenue", the answer is not a fixed count and Top N is the wrong shape of tool.

What Are the Limitations of Top N Filtering?

There are three limits worth knowing before you build a report around this filter, and the first one has caused more quiet reporting errors than the other two combined.

Totals Reflect Only the Filtered Rows

Once the filter is applied, the visual's total row is the total of the top ten, not the total of everything. A table showing ten products with a 2.4 million total looks authoritative and is not the company's revenue. If the reader needs both numbers, put the full total in a separate card visual that does not carry the Top N filter, and label both clearly.

There is no built-in "Other" bucket either. Power BI will not add an eleventh row collecting the remaining 390 products. If you need that, you need DAX.

Ties Are Resolved Silently

If three products are tied at the tenth position, the visual still shows exactly ten rows and does not tell you a tie occurred. One of the tied items is in and the others are out, and which one wins is not something you control from the filter card.

This is a real difference from the DAX approach. Microsoft's documentation for the TOPN function states that when there is a tie in the ordering values at the N-th row, all tied rows are returned, so TOPN can return more than N rows. The visual filter does not behave that way. If tie handling matters, which it does for anything audited or incentive-linked, use the DAX route and decide the tiebreak deliberately.

N Has to Be a Fixed Number

The Show items box takes a literal integer. You cannot bind it to a parameter, a slicer or a measure, so you cannot ship a report where the reader chooses between top 5, top 10 and top 25 from a dropdown using this filter alone. Doing that requires a numeric range parameter and a measure that reads it, which is a different build.

Guessing which pages deserve your attention? Get a free SEO audit and find out which ones are actually close to ranking.

Top N Filter vs the DAX Alternatives

When the built-in filter is not enough, there are three standard ways to do ranking in DAX. Each solves a different one of the limitations above.

The TOPN Function

TOPN returns a table of the top N rows, which makes it the natural building block inside an aggregation. Its syntax is TOPN(<N_Value>, <Table>, <OrderBy_Expression>, [<Order>]), and a measure that sums revenue for the top ten products looks like SUMX(TOPN(10, SUMMARIZE(Sales, Sales[Product], "Rev", SUM(Sales[Amount])), [Rev], DESC), [Rev]).

Two documented behaviours are worth carrying in your head. It returns all tied rows at the boundary, so the result can exceed N. And it guarantees no particular sort order in its output, so if you are displaying the result you still have to sort it yourself. It is also unsupported in DirectQuery mode when used in calculated columns or row-level security rules.

The RANKX Function

RANKX gives every row a rank number rather than returning a shortened table, which is what you want when the rank itself should be visible or when the cutoff needs to be flexible. A measure like RANKX(ALL(Products[Product]), [Total Revenue]) produces a position for each product, and you then filter the visual to ranks less than or equal to ten, or compare the rank against a parameter the reader controls.

This is the approach that unlocks the dynamic-N report the built-in filter cannot deliver, and it lets you show the rank as a column, which readers almost always like. The cost is that you have to be deliberate about the first argument: get the ALL or ALLSELECTED choice wrong and your ranking will reset inside every row of the visual and return 1 everywhere.

A Calculated Table or an Other Bucket

If the requirement is "top ten plus everything else as one row", neither the filter nor a simple measure gets you there. You build a calculated column or table that labels each member as either its own name or "Other" based on its rank, then group by that label. It is more work and it is the only approach that produces a complete picture in a single visual, which is why it keeps appearing in executive summaries.

For a grounding in writing these measures properly, our guide to how to write DAX functions in Power BI covers filter context and the handful of functions that do most of the work.

Final Thoughts

The built-in Top N filter is the right default. It takes fifteen seconds, it respects the filter context so slicers behave the way readers expect, and it covers the plain "show me the leaders" request completely.

Reach for DAX when one of three things is true: the totals need to represent the full dataset, ties have to be handled explicitly, or the reader needs to choose how many rows to see. Those are the real boundaries of the filter, and all three are predictable enough that you can tell which tool you need before you start building rather than halfway through.

The habit worth forming is the one from the top of this guide. Before you ship a visual showing the top ten of anything, look at its total and ask whether a reader would read that number as the whole business. If they would, either relabel it or move it somewhere that is not filtered.

Ready to find the gaps in your own reporting and rankings? Start with a free SEO audit and work from data instead of guesswork.