This message was deleted.

Hello! It looks like your message is incomplete—could you please provide your full question or let me know how I can assist you regarding GoodData? I'm here to help with any GoodData-related queries.

When downloading a dashboard as a PDF, if it includes a table, the exported file doesn’t capture the full table, a few columns on the right keep getting cut off. We’ve tried adjusting the table size and increasing the right margin, but it still happens. Is there a way to fix this so the full table can be exported without anything being cut off?

To use segmentation in a scatter plot in GoodData Cloud, the attribute must already exist in your data model. This means the segmentation logic (e.g., “Green” or “Red” based on growth) must be represented as a column in your source data or created via a SQL Dataset. GoodData Cloud does not support creating “computed attributes” or dynamic attributes based on metric logic (such as a “Green/Red” label based on a metric value) within the platform itself. How the attribute approach works: • An attribute in GoodData is a logical object representing a non-measurable descriptor, such as a category or label, that provides context to your data. Attributes are used to break down metrics and can be used for grouping, filtering, or segmenting data in visualizations. • In your data model, each attribute is mapped to a column in your database or data source. For example, if you want to segment by growth status, you would have a column in your data source (or a computed column in a SQL Dataset) with values like “Green” or “Red” for each record. • Once this attribute is present in your data model, you can use it in the “Segment by” field of a scatter plot or other visualizations. Relevant documentation: • The concept of attributes and how they are used in datasets is explained here: Attribute. • For creating new attributes using SQL logic, you can use SQL Datasets, which allow you to define a dataset based on a custom SQL query. This is useful for creating computed columns for segmentation: Create SQL Datasets. Summary: To segment your scatter plot, ensure your segmentation logic is represented as an attribute in your data model—either as a column in your source data or as a computed column in a SQL Dataset. This attribute can then be used for segmentation in your visualizations. If you need to create such an attribute based on logic, SQL Datasets are the recommended approach in GoodData Cloud.

Thanks Michael, another question, is there a way to sort the months in actual month order instead of alphabetically? Our financial year kicks off in July, so having it come out as ‘Apr, Aug…’ looks pretty silly, it should be Jul, Aug ... The other question I’ve got is about DISTINCT. Since GoodData doesn’t support DISTINCT directly, we’re using a BY clause to get unique values and then summing them. That part works. But as soon as we apply

FOR PREVIOUS

, the LY result is identical to this year.

FOR PREVIOUS(BUS_DATE.year)

works on our other metrics, just not this one. Any idea why that’s happening and how we sort it out?”

By default, months are sorted alphabetically unless you use a date attribute that encodes the chronological order (e.g., a “Month of Year” attribute with a sort label or a custom fiscal calendar). You can try and use the “Sort by” option in the modeler to select a column for an attribute that will be used for ordering attribute values. If you have a column in your data that represents the fiscal month order (e.g., 1 for July, 2 for August, etc.), you can set this as the sort label for your month attribute. This will allow months to be displayed in the correct fiscal order rather than alphabetically. Note that this approach requires your data model to include such a column and for it to be mapped appropriately in the modeler. The good news if though that we will soon releasing flexible fiscal calendar support, beginning with month-offset capability - enabling analytics that align with organizations whose fiscal year differs from the standard Gregorian calendar. Please follow our release notes to stay upto date with all the new features we will be releasing: https://www.gooddata.com/docs/cloud/whats-new-cloud/ Regarding the DISTINCT question: Because the metric uses a DISTINCT-style (nested aggregation with a BY clause), the link to the date dimension is not straightforward for the time-shift function. As a result,

FOR PREVIOUS

cannot correctly move the calculation one year back and returns the same value as the current year. I would suggest to try and apply

FOR PREVIOUS

to a simpler base metric and then aggregate it, or you might have to adjust the data model/metric to use the date dimension directly that the fact is connected to.

Thanks Michael, Hope we can get that feature down the track. I’ve got another question about summing distinct values. I’m trying to calculate YTD vic_revenue. I’ve added an lga_curr_ytd flag in the table so I can mark which rows fall in the current financial year (our FY starts on 1 July). The issue is that vic_revenue is the same value across all rows for the month, for example, there might be 20 rows all showing the same revenue figure. What I want to do in GoodData is sum the distinct vic_revenue values. How can I set this up? I've tried few metrics but I'm not getting the correct outcome yet.

Hi Cathy, For the YTD VIC revenue, since

vic_revenue

repeats across many rows for the same month, you need to make sure each month is only counted once. GoodData doesn’t have a direct

SUM DISTINCT

, but you should be able to achieve the same result using a nested metric:

-- YTD VIC Revenue (sum of distinct monthly values)
SELECT SUM(
  SELECT MAX( VIC REVENUE {fact/VGCCC_VENUE_MONTHLY_AGG.LGA_REVENUE} )
  BY Month - Month/Year {label/MONTH.month}
  WHERE LGA CURR YTD {label/lga_curr_ytd} = "Y"
)

The inner part groups the data by month and takes

MAX()

of

vic_revenue

— since the revenue is identical for all rows of that month, this gives one value per month. • The outer

SUM()

then adds those monthly values together to get the correct YTD total without double-counting rows.

Thanks Julius, it works. How can I create a map in GoodData for Victoria showing our venues, with the bubble size representing YOY growth? We’ve got latitude and longitude in our data. I’m used to doing this in Tableau—it’s pretty easy there. But it looks like there’s no X and Y here. What’s the ‘Location’ for? Is it a name, an address, or some combination of latitude and longitude?

Hi Cathy, happy to help here. Before building the map, you’ll first need to prepare your data and Logical Data Model to support geo visualizations, In GoodData, the map expects a Location attribute, which represents the logical entity you want to place on the map (for example, a venue). That location can be resolved either by built-in geocoding (city, state, or country names) or — as in your case — by explicitly providing latitude and longitude. Since you already have latitude and longitude, the right approach is to use a Geo Pin map. Once your Year-over-Year (YOY) growth is defined as a metric, it can be used directly to control the bubble size on the map. I am referring our documentation where you can see the correct setup. Please give it a try and let us know how it goes.

Thanks Mauricio, for the quick reply. I’ve set up labels for latitude and longitude in the LDM, but I’m getting some errors and can’t save it. Any idea what might be wrong?

Hi Cathy, I am not sure if I understood well your definition of the attributes and labels. The error message as you can see states:

message=Attribute 'latitude' is missing additional LONGITUDE/LATITUDE label

But form your second screenshot, it seems you defined twice the latitude and longitude as independent attribute and label). As example, you can see the attached screenshot how it is working for me, on my end, there I defined the location as label and then the corresponding latitude and longitude.

This works, thanks! But how do I make the map show just certain states instead of the whole world in the dashboard? I've filtered map to Australia in the view.

Hi Cathy, as mentioned in our documentation: To set a specific part of the map to be seen when you open the visualisation, go to Configuration > Map > Default viewport and select a part of the map. https://www.gooddata.com/docs/cloud/create-visualizations/visualization-types/geo-charts/ That setting only controls what part of the map is shown when the insight opens, In this case, you’ll need to use filters to restrict what appears: Use filters (for example, a

State

filter) so that only points from the desired states are plotted.

Hi Michael,I’ve set up a state filter on the dashboard and set the default viewport to Australia (that’s the smallest range available), but the map is still displaying as shown in the screenshot. Is there any other setting I can use to make the map only display the states selected in the filter?

Hello Cathy, I have been investigating this on my end and found that the display is directly affected by the size and "frame" of the insight on your dashboard (plus Australia is on the edge of the map layout). When I replicated the exact shape of your insight, I encountered the same behavior when filtering for Australia. However, when I resized the insight to a more "square" shape, the map appeared more zoomed-in (as shown in the first attached image). I also tested this with Europe, which seems to scale better within the current frame, this also can be because Europe is more in the center of the map layout (second image). The behavior you are seeing is due to a current platform limitation involving a static and default zoom level for maps. The good news is that our product team is already aware of this. A "configurable zoom level" feature is on our roadmap for next year. While we don't have a specific ETA yet, I recommend keeping an eye on our Releases Page for updates. Please let us know if this clarifies the situation or if you have any further questions.

Thanks Mauricio, hopefully we can get the feature soon. Cheers

Hello Cathy, I believe this is currently not possible. Attributes placed in the Columns section of a pivot table cannot be sorted ascending or descending. As a workaround, similar as it was mentioned to you, you can create an additional helper attribute (or column) that defines the desired order, but this time another order — for example, July = 12, August = 11, and so on. You can then use this helper attribute to control the column order indirectly by sorting on it. This approach allows you to achieve the required ordering even though direct column sorting isn’t supported.

Hi Mauricio, I think this might be getting a bit mixed up. I’ve already defined a financial month number to ensure the months are ordered correctly (Jul → Jun). What I’m trying to do here is sort the financial year and month number in descending order, like FY 2026 → 2025 → 2024, and within each year 12 → 1, for fy2026 4->1. In the current view (see screenshot below), there doesn’t seem to be any option to sort the pivot table columns. Previously, we were able to apply sorting, but that option no longer appears to be available.

Hello Cathy, I understand now... but are you sure the 'sort' screenshot is for Pivot tables? As you can see in our documentation, the sort option shows up mainly for charts, rather than Pivot tables. I will reference in here for your convenience:

Sorting is available for:

• bar charts

• column charts

• area charts

• combo charts

• bullet charts

• pie charts

• donut charts

• heatmaps

You can sort data:

• Alphabetically

• Numerically

• Chronologically

Hi Mauricio, I've seen the sort before, but I can't find it now.

Hi Cathy, that option is not visible for Table reports - But if you want to sort the table, you need to click on the icon highlighted below: