I have a table in Tableau that is based on a MTD calculation. It gives multiple values for different departs on each day of the month. This table is linked to another sheet that I click on to choose the value in the table that I want to see.
Right now I only highlight that value in the table. I want to be able to pull one specific value out and display by itself in the dashboard close to the chart.
How do you pull a value from a table and display.
Related
I have a Dashboard with different visuals.
Data is made up of different values for insurance companies.
I want my slicers/filters to not filter all data, but to only highlight the chosen company.
For example, in my slicer I choose the insurance ABN.
Instead of showing me the value for ABN only in my visuals, I want all other values to still be visible and ABN's value to be highlighted in the visuals.
Does anyone know how to do this?
You can use conditional formatting to achieve this. Lets say that we will change the background color to "highlight" a row (or cells, to be precise).
First, we need a slicer, which will not filter our data. We can do this by duplicating our source table, removing the unnecessary columns and making sure there is no relationship between the source and the duplicate. So if we have a source table, named Table, like this:
Right click on it and select Duplicate:
Then right click the title of the column you want to keep and select Remove Other Columns to get a list of company names only (you may also remove the duplicates, but it's not required). Then in the model delete the relation between both tables:
Now you can place a table showing company name and sales from your data source, and a slicer for company name from the duplicate table. At this point selecting values in the slicer should not affect the table.
Now you need to capture the value of the slicer and use it in a measure, which will determine should current row be highlighted or not. You can use SELECTEDVALUE for that, but note that it will give you a value only if there is a one selected in the slicer. If you want to support highlighting of more than one company, it gets a bit more complicated.
Make a new measure in your source table, like this:
Measure = IF(HASONEVALUE('Table (2)'[Company name]);
IF(SELECTEDVALUE('Table (2)'[Company name]) = MAX('Table'[Company name]); 1; 0);
IF(ISFILTERED('Table (2)'[Company name]) && COUNTROWS(FILTER('Table (2)'; 'Table (2)'[Company name] = MAX('Table'[Company name]))); 1; 0))
In case there is only one value selected in the slicer (see HASONEVALUE), then our measure will return 1 (highlight) or 0 (don't), comparing it with the current row.
Otherwise (i.e. there is no selection in the slicer, or there are 2 or more companies selected), then we will look at the filtered list of companies (Table (2)) - if it contains current row, then 1 (highlight), otherwise 0 (don't). But we will also handle the case, where there is no value selected in the slicer. In this case the list will contain all the companies, i.e. all rows will be highlighted. Here comes ISFILTERED. And at the end, if the list is filtered and current row exists in the filtered list, then 1 (highlight), otherwise 0 (don't).
Now, you need to use this measure to change the background of the column - right click each column in your table and select Conditional formatting -> Background color:
Then format by rules, where Measure >= 1 like this:
Now, when there is no selection in the slicer, there are no rows highlighted in the table:
If you select one company, it is highlighted:
It also work if there are multiple companies selected:
Thank you Andrey for your step-by-step explanation which as been incredible helpful. I'd like to follow up with a further question, particularly regarding the comment below.
"You can use SELECTEDVALUE for that, but note that it will give you a
value only if there is a one selected in the slicer. If you want to
support highlighting of more than one company, it gets a bit more
complicated."
In my model, I've linked a third table (Table (3)) to Table (2) with a many to one relationship with Table (2). Therefore when I click on Table (3), it will filter Table (2), which acts as a slicer for Table (1).
When only 1 value is filtered in Table (2), it conditionally formats the cells in Table (1). However, when more than 1 value is filtered in Table (2), conditional formatting fails.
As I'm looking to avoid manually selecting multiple values in the slicer (Table (2)), I was wondering if there's a workaround for SELECTEDVALUE such that it is able to conditionally format when I filter more than 1 value in Table (2).
See image below. Warehouse and Customer are dropdown lists populated via query. Year is a text field. The two tables display data from datasets driven by the report parameters. Is there a way to repeat the two tables based on each member of the Customer dropdown? Preferbly with a pagebreak after the 2nd table.
I normally do this using subreports. You could either create a single subreport that contains both tables or individual subreports. The individual approach might help with page breaks etc so that's the way I'd go.
Step 1: Create a report for your first table.
As you don't state what each table does, I'll make some up for the sake of illustration.
The key is to create a subreport that displays just the info you need in a single table. So in your case this might mean we only need to pass in a single parameter, CustomerID. You might need to pass in more such as Warehouse but I don't know...
In my made up scenario, let's assume the report shows customer contacts so we create a subreport (let's call it subCustomerConacts). It has a single parameter pCustomerID and a single dataset dsCustomerContacts. The query might looks something like SELECT * FROM CustomerContacts WHERE CustomerID = #pCustomerID. Add whatever tables/textbox etc you need to display your data.
Test this subreport works by manually typing in a CustomerID
Step 2: Create a report for your second table.
Do exactly the same again, creating a new subreport. Let call this subCustomerOrders. Repeat as above until you end up with another report that can display order details (or whatever you need).
Finally, create you main report.
This is basically what you have described in you question in terms of parameters etc.
Now to add this bit that will call you subreports.
Create a dataset (let's call it dsCustomerLoop) that contains each customer from your parameter something like SELECT DISTINCT CustomerID FROM myCustomers WHERE CustomerID IN (#myCustomerParameter)
Add a table to your report, 1 column wide and stretch it so it's wide enough to accommodate you subreports.
Set the dataset to point to dsCustomerLoop
Right-Click the cell in the detail row and do "Insert Row -> Inside Group - Below". You should not have two detail rows.
Next, right click the top detail row and do "Insert -> Subreport"
Right-Click the newly inserted subreport control and choose "properties".
Choose your first SubReport form the drop-down list
Click parameters on the left,
Click "Add" and select the CusomterID parmeter, set it's value to the CustomerID field.
Repeat this process on the seconds row, choosing your seconds subreport.
You may want to also add a 3rd row to the table, you could insert a rectangle into this with page breaks set to force a new page after each seconds subreport.
That's It. When the report runs it will produce two rows per customer, each row containing a subreport.
I hope this is clear enough, I've rushed through it a bit but if anything is unclear, let me know and I'll provide a clearer solution.
I have consulted this question - OBIEE Moving Average (Mavg) for 4 weeks on Pivot Table - in order to understand how to do a moving average in OBIEE. However, I am having trouble calculating that item in a pivot table.
In the pivot table view I thought I would just select New Calculated Item and create my moving average function for the pivoted values. Yet, in the "Values From" drop down menu the items I want to average are not present. I believe this may be because they are not extracted data values but previously calculated values from the data (in the original table's columns). I tried selecting "treat as attribute column" but this failed as well. How can I created a moving average within the pivot table for a column that was calculated from the original data?
Looks similar to this:
Pivot Table
ID Value
01 45
02 54
03 65
... ...
Where Value is Amount Sold / Days. And both Amount Sold and Days are stored in the original data table.
Go to the criteria tab and pull in a measure (any measure will do). Now edit that measure's column formula to your formula of Amount Sold / Days. I would rename it as well.
When you click over to the results tab, you will see that column added to all views. Just edit each view and remove the column from the views you do not want to see it in. You can also replace your existing value column in your pivot table (if it was present) with this new calculated column.
Pivot table calculated items are typically used in conjunction with columns that are already in the pivot table.
Does any one know If there are any best-practices for editing/adding data in UI table while its data is filtered?
Example 1:
Suppose we have a table with two options: add new record and edit selected one. Moreover, the table has an option to filter data over column A.
Now, if the table is filtered by filtering column A with value '1' and I want to add a new record with value in column A that matches filter requirements, what should happen:
The table should refresh and display filtered records with selected newly added record.
The table should reset filter and show all records witch new one selected.
The table should do nothing and display filtered records as they were. Newly added record will be displayed when the filter resets.
For me intuitively the best solution is number 1. But then how to solve the problem in example number 2:
Example 2:
If the table is filtered by filtering column A with value '1' and I want to add a new record that in column A has value '2' what should happen:
The table should display filtered records with selected newly added record despite it does not matche the filter.
The table should reset the filter, and all records should be displayed with new one selected.
The table should do nothing and display filtered records as they were.
The same story is when we have filtered records, and in selected record we want to edit value upon which the filtering took place. Does the edited record should than disapear or filtre should be reset?
Or maybe the best way is to disable add/edit operations while filter is on?
I don't know if there is any best practices about it but I have also encountered the problem before. Two different solutions as I came up with:
Edited/Inserted record should not be filtered until the next time a filter is applied or filter is reset. The record should also be shown different (i.e. darker background color, or an icon, or tooltip) than others implying it was edited and is not being filtered.
The record should be left in focus after being edited or inserted. As soon as it loses focus, filter should be applied to it. The ideal solution is if the record is filtered out, it shouldn't immediately go out of vision. For instance it may go invisible with an animation.
I just checked the google docs and libre office, both of them just display the new record regardless of the filter. You need to re-apply the filter to hide them from view (in both cases the column "A" is filtered by "value 1"):
I have a CSV file that looks like this. It is in matrix form where cell A1 is empty. I want to visualize this in Tableau like this
a1,a2,a3
a1,1,0.2,0.3
a2,0.3,1,0.5
a3,0.6,0.7,1
How can I build this in tableau to show a matrix?
I tried putting measure name in both the row and column field and measure values in the marks text box but i ended up with diagonal values which isn't what i want.
First structure your data in the form of a table.
A well structured CSV file for Tableau (and most tools) has a single header row that labels each column (so skipping the first cell in the header row is a problem). Then each data row that follows contains only values in each field separated by columns (so there should be no leading row labels)
Order is not significant in a relational table. The rows in a table form a set, not a matrix.
So your CSV should look something closer to
field_name_1,field_name_2,field_name_3
1,0.2,0.3
0.3,1,0.5
0.6,0.7,1
Normally, the prime use of Tableau is display an aggregated summary of the data, not to display each individual data row. So you can put measure_names on columns and measure_values on the text or label shelf to get a summary of each measure (by default a sum)
If all you want to do is to view the underlying data rows in a table for diagnostics, the easiest way is to use the view data button or menu command.
If you want to build a visualization that displays all the data (without summarizing it) in a table, you have two ways to go. The simplest is if you have a unique primary key in your data, you can place it on the row shelf to get a row in your viz for each row in your data.
If your data rows do not have a primary key, you can turn off aggregate measures from the Analysis menu, but then you'll see multiple values in each table cell. Then to get a row in your viz for each row in your data, put the calculation index() on the rows shelf and change it to discrete.