# Resolve tables Every tile that shows data counts through one table, called its resolve table. By default, it's the parent table of the tile's dimensions. When a dashboard has tiles built from different tables, Orbit keeps your filters for each table separate, and combines them into the right filter for each tile. ## How filters work across tables When you select categories on tiles, the user filters panel groups your filters by table. Each tile then applies every block through its own table. > [!NOTE] > > **Filters on two tables** > > A dashboard has a **Booking Destination** chart and a **Booking Date** chart, both from the **Bookings** table, and a **Policy Cover** chart from the **Policies** table. Select **Australia** on Booking Destination, **2018** on Booking Date, and **Individual** on Policy Cover. The user filters panel shows a **Bookings** block with two filters and a **Policies** block with one. ## See the filter a tile uses **Underlying Filter** shows the full filter for a tile, including filters from other tables, written out for the table the tile counts through. To see a tile's filter, expand the **option menu** on the tile and select **Underlying Filter**. ## Filter on pairs of values When you select categories on a [multi dimension chart](), each selection pairs a category from each dimension, such as a year and a destination. The pairs stay together in one filter block, so you can build filters that single dimension tiles can't. > [!NOTE] > > **France in 2017 or Germany in 2018** > > To find people who booked France in 2017 or Germany in 2018, you could select **France** and **Germany** on the **Booking Destination** chart, and **2017** and **2018** on the **Booking Date** chart. That gives two separate filters, which also select people who booked France in 2018 or Germany in 2017. > > Instead, use a multi dimension chart with **Booking Date** and **Destination** as its dimensions. Select the **France** column in **2017** and the **Germany** column in **2018**, and add them to your filters. One **Booking Date & Destination** block holds the pairs **2017 & France** and **2018 & Germany**, and selects only the people you wanted. The count drops from 305,392 people to 164,937. ## When to change Select Via You can make a tile count through a different table with its **Select Via** setting. Changing the table changes what a filter means for the tile, so the same filter can answer a different question. > [!NOTE] > > **Booking destinations through People** > > A **Booking Destinations** chart counts bookings by destination. With **Select Via** on **Auto (Bookings)**, filtering the dashboard to **United States** leaves only the United States bookings on the chart. > > Set **Select Via** to **People** and the same filter selects the people who booked a trip to the United States. The chart then shows every destination those people have booked, not only the United States. ## Change the Select Via table **Select Via** is in the **Data Settings** of every tile type except text and profile tiles. To change the table a tile counts through: 1. With your dashboard in edit mode, expand the **option menu** on your tile and select **Edit Tile**. 2. Expand **Data Settings**. 3. Select a table from the **Select Via** dropdown. **Auto** uses the parent table of your dimensions, shown in brackets, for example **Auto (Bookings)**. 4. Select **Apply**, then **Save** the dashboard. The tile now counts through the table you chose whenever a filter applies. > [!TIP] > > Add a [note]() to the tile to tell viewers which table it counts through, so its numbers don't surprise them. ## Include records with no matches When you choose a table above the tile's own table, such as **People** for a tile that counts bookings, **Data Settings** shows two more toggles: - **Return All Values**: Includes records from the **Select Via** table that have no matching records, for example people with no bookings - **Apply Filter To**, followed by the tile's table, for example **Apply Filter To Bookings**